This is the accessible text file for GAO report number GAO-09-56
entitled 'Highway Safety: Foresight Issues Challenge DOT's Efforts to
Assess and Respond to New Technology-Based Trends' which was released
on October 6, 2008.
This text file was formatted by the U.S. Government Accountability
Office (GAO) to be accessible to users with visual impairments, as part
of a longer term project to improve GAO products' accessibility. Every
attempt has been made to maintain the structural and data integrity of
the original printed product. Accessibility features, such as text
descriptions of tables, consecutively numbered footnotes placed at the
end of the file, and the text of agency comment letters, are provided
but may not exactly duplicate the presentation or format of the printed
version. The portable document format (PDF) file is an exact electronic
replica of the printed version. We welcome your feedback. Please E-mail
your comments regarding the contents or accessibility features of this
document to Webmaster@gao.gov.
This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed
in its entirety without further permission from GAO. Because this work
may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this
material separately.
United States Government Accountability Office:
GAO:
Report to the Chairman, Committee on Commerce, Science, and
Transportation, U.S. Senate:
October 2008:
Highway Safety:
Foresight Issues Challenge DOT's Efforts to Assess and Respond to New
Technology-Based Trends:
GAO-09-56:
GAO Highlights:
Highlights of GAO-09-56, a report to the Chairman, Committee on
Commerce, Science, and Transportation, U.S. Senate.
Why GAO Did This Study:
Fatalities on U.S. roads now total over 40,000 each year. Future
reductions may require the Department of Transportation (DOT) to
address new trends such as evolving crash-avoidance technologies and
rapidly changing electronic devices that may distract drivers who use
them on the road. (See figure.) GAO was asked to examine how DOT is
addressing fast-moving trends such as these. This report examines how
DOT is (1) deciding on responses to the crash avoidance and electronic
distractions trends—given available evidence and uncertainties; (2)
developing new evidence on these trends’ safety impacts; and (3)
communicating with the Congress about these and other trends and
related issues. To conduct this study, GAO analyzed DOT reports, peer-
reviewed literature, and other documents; interviewed DOT officials and
staff; and interviewed over 30 experts.
What GAO Found:
New fast-moving technology-based trends are characterized by
uncertainties, and the main criteria that DOT’s National Highway Safety
Administration (NHTSA) officials use in deciding how to respond—
quantitative evidence that a sizable problem exists and knowledge of a
promising countermeasure—do not address uncertainty. One technology-
based trend presents potential opportunities to improve future safety:
evolving crash avoidance technologies. With somewhat limited data on
actual safety benefits, NHTSA is pursuing such opportunities by, for
example, providing consumer information about new car technologies
designed to help avoid some crashes. A different trend represents a new
threat to safety: rapidly changing and proliferating electronic driver
distractions. Although NHTSA is conducting studies to understand this
trend’s nature and scope, it is not self-initiating actions or research
designed specifically to counter new distractions, citing a lack of
evidence that these are as significant a problem as, for example,
failure to use seatbelts. Literature and experts suggest alternative
approaches to decision-making, such as anticipatory risk management and
expansion of networks, which might help with decisions about
investments to shape or counter fast-moving trends.
DOT also faces challenges in developing additional, higher quality or
more timely evidence on the changing sizes of the safety impacts of
such trends—despite attempting to obtain appropriate data through both
long-standing and new methods. For example, analyses of existing crash
datasets produce valid comparisons of crashes in cars with and without
new technologies, but such analyses require years of accumulated
results and thus cannot keep pace with a fast-moving trend. Developing
more timely, high-quality evidence would (1) improve evaluations of new
safety technologies’ benefits and (2) identify the level of threat
presented by evolving driver distractions—thus reducing uncertainty and
supporting decisions. Innovative approaches, such as data collection
that uses emerging technologies for wireless transfer of crash data or
new analysis techniques, might help provide more timely, high-quality
evidence on the impacts of trends and how these change over time.
Figure: Example of Type of Crash That New Safety Technology Is Designed
to Help Drivers Avoid and Dashboard with Devices, Such As Cell Phones,
That Might Distract Drivers (Illustrations):
[Refer to PDF for image]
Source: GAO and ProClip (portable device holders).
[End of figure]
DOT currently communicates some relevant information to the Congress on
emerging trends but these communications are not designed to provide a
long-term view of highway safety, including trends such as evolving
crash avoidance technologies and rapidly changing electronic driver
distractions—and their implications for the years ahead—together with
timely updates. DOT has not synthesized the results of its work for the
Congress to look at how overall trends will impact highway safety in
2020 and beyond. Some of DOT’s own practices and other models from the
United States and abroad might provide improved strategies for
communication.
What GAO Recommends:
GAO recommends that DOT (1) develop an approach to guide decision-
making on new, fast-moving trends that can affect highway safety; (2)
evaluate whether new data systems and analytic techniques are needed to
provide information on such trends; and (3) employ specific strategies
and schedules in communicating with the Congress about these and other
trends. DOT disagreed with the first of these and did not comment on
the other two. GAO continues to recommend all three.
To view the full product, click on [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-09-56]. For more information, contact Nancy R. Kingsbury
at 202-512-2700, kingsburyn@gao.gov, or Katherine A. Siggerud at 202-
512-2834, siggerudk@gao.gov.
[End of section]
Contents:
Letter:
Results in Brief:
Background:
Uncertainty about New Opportunities and New Threats from High-
Clockspeed Trends Challenges DOT Decision Makers:
DOT Faces the Challenge of Devising Timely Measures of Change over
Time:
DOT Faces the Challenge of How and When to Communicate Information on
Trends to the Congress:
Conclusions:
Recommendations for Executive Action:
Agency Comments:
Appendix I: Exercising Foresight: Three Agency Activities and Related
Challenges:
Appendix II: Private Sector, University, and Other Experts We
Interviewed:
Appendix III: Review of Studies of Driver Phoning and Highway Safety:
Appendix IV: Trends for Vulnerable Road-User Groups:
Appendix V: Illustrations of In-Vehicle Screens for Safety:
Appendix VI: Information on Selected NCSA Datasets:
Appendix VII: In-Vehicle Crash Avoidance Technologies and New NCAP:
Appendix VIII: Unintended Consequences:
Appendix IX: Literature Review's Suggestion: Evaluating State Laws:
Appendix X: Comments from the Department of Transportation:
Appendix XI: GAO Contacts and Staff Acknowledgments:
References:
Tables:
Table 1: Methods for Assessing the Safety Impacts of Evolving Post-ESC
Crash Avoidance Technologies Rated by Three Key Characteristics:
Table 2: Methods for Assessing the Safety Impacts of Evolving
Electronic Driver Distractions Rated by Three Key Characteristics:
Table 3: Examples of Foresight across DOT:
Table 4: Impact of Driver Phoning on Crash Risk: 13 Primary Studies:
Table 5: Technical Adequacy Descriptors: Primary Studies of Crash Risk
by Design Category:
Table 6: Selected NCSA Datasets:
Figures:
Figure 1: Potential Consequences of Falling behind a Damaging Trend
That Could Be Effectively Countered by Early Action: Illustration from
Foresight Literature:
Figure 2: Exercising Foresight by Addressing High-Clockspeed Trends:
Three Challenging Agency Activities:
Figure 3: Actual Overall Fatality Rates and Number of Fatalities, 1975
through 2006, and Rates Projected through 2016:
Figure 4: Evolving In-Vehicle Crash Avoidance Technologies:
Figure 5: Three Major Types of Crash Avoidance Applications:
Figure 6: The Evolution and Spread of Portable Electronic Devices in
the United States:
Figure 7: Overall Minutes of Cellular Use in the United States, 1997-
2007:
Figure 8: Touch Screens and Virtual Screens:
Figure 9: Foresight Reporting Framework for High-Clockspeed Trends:
Figure 10: Exercising Foresight by Addressing High-Clockspeed Trends,
with Deciding and Responding Highlighted:
Figure 11: Foresight Reporting Framework: Post-ESC In-Vehicle Crash
Avoidance Technologies:
Figure 12: How VII Would Operate:
Figure 13: Foresight Reporting Framework: Evolving Electronic Driver
Distractions:
Figure 14: Dashboard Mounts and Helmet Equipment for Phoning:
Figure 15: Exercising Foresight by Addressing High-Clockspeed Trends,
with Evidence Development Highlighted:
Figure 16: Exercising Foresight by Addressing High-Clockspeed Trends,
with Agency Communication Highlighted:
Figure 17: Small Cars as a Percentage of Passenger Vehicle Sales,
January 2003 to June 2008:
Figure 18: Screen with V2V Icon Warning of a Stopped Vehicle in the
Road Ahead:
Figure 19: Backup Camera Screen (Activated When Car Is in Reverse):
Figure 20: Unintended Consequences of the Interaction of Multiple
Trends:
Abbreviations:
ACAT: Advanced Crash Avoidance Technologies:
CEA: Consumer Electronics Association:
CICAS: Cooperative Intersection Collision Avoidance System:
DOT: Department of Transportation:
ESC: electronic stability control:
Euro NCAP: European New Car Assessment Programme:
EVSCA: Effectiveness of Vehicle Safety Communications Applications:
FARS: Fatality Analysis Reporting System:
FCW: forward collision warning:
FHWA: Federal Highway Administration:
FMCSA: Federal Motor Carrier Safety Administration:
FOT: field operational test:
GDL: graduated drivers licensing:
GHSA: Governors Highway Safety Association:
GPRA: Government Performance and Results Act of 1993:
GPS: global positioning system:
LDW: lane departure warning:
MMIRE: Model Minimum Inventory of Road Elements:
NASS CDS: National Accident Sampling System's Crashworthiness Data
System:
NCAP: New Car Assessment Program:
NCSA: National Center for Statistics and Analysis:
NHTSA: National Highway Traffic Safety Administration:
NOPUS: National Occupant Protection Use Survey:
NTSB: National Transportation Safety Board:
RITA: Research and Innovative Technology Administration:
SHRP 2: Second Strategic Highway Research Program:
TIA: Telecommunications Industry Association:
TRB: Transportation Research Board:
V2V: vehicle-to-vehicle communications:
VII: Vehicle Infrastructure Integration:
VMT: vehicle miles traveled:
[End of section]
United States Government Accountability Office:
Washington, DC 20548:
October 3, 2008:
The Honorable Daniel Inouye:
Chairman:
Committee on Commerce, Science, and Transportation:
United States Senate:
Dear Mr. Chairman:
Reducing future highway fatality rates (defined as the number of road
fatalities per vehicle mile traveled) is an important national
objective. Traffic deaths on U.S. roads total over 40,000 each year,
and although past technology developments and government actions
substantially reduced fatality rates, earlier progress has now slowed.
Projections from the Department of Transportation (DOT) suggest that as
many as 500,000 deaths could occur on U.S. roads between now and 2020,
unless vehicle miles traveled (VMT) substantially decrease or progress
in reducing fatality rates improves.
A variety of ongoing or anticipated trends--ranging from the aging of
the U.S. population and the increasing use of motorcycles to fast-paced
technology-based trends--have the potential to transform the highway
safety landscape, for better or for worse. Future progress in reducing
fatality rates may depend on DOT's ability to exercise foresight by
addressing potentially significant but somewhat uncertain
trends.[Footnote 1] The most challenging of these may be technology-
based trends that proceed at a high clockspeed, that is, (1) a faster
pace than trends DOT has dealt with previously, especially with respect
to new products that can affect safety, or (2) a quantitative rate of
change that is either exponential or exhibits a pattern of doubling or
tripling within 3 or 4 years, possibly on a repeated basis.[Footnote 2]
Our 21st Century Challenges report raises the issue of whether federal
agencies are poised to address fast-paced technology-based changes (GAO
2005a). This and other analyses suggest that unless agencies and the
Congress can stay abreast of technological changes, they may find
themselves "in a constant catch-up position and lose the capacity to
shape outcomes" (Rejeski and Wobig 2002, 15).[Footnote 3] And foresight
literature illustrates the potential future consequences of falling
behind a damaging trend that could be countered by early action. As
indicated by figure 1, early intervention in this situation can be
effective, and delaying intervention can result in negative outcomes.
Figure 1: Potential Consequences of Falling behind a Damaging Trend
That Could Be Effectively Countered by Early Action: Illustration from
Foresight Literature:
[Refer to PDF for image]
This figure is an illustration of potential consequences of falling
behind a damaging trend that could be effectively countered by early
action in a graph format. The vertical axis of the graph represents
damage, escalating from minimal to reversable to irreversable and very
costly. The horizontal axis of the graph represents time periods
current, near-term future (somewhat uncertain) and longer-term future
(very uncertain).
Line are depicted on the graph representing the following:
Current, potentially damaging trend;
Projected, without intervention;
Projected, with early intervention;
Projected, with late intervention.
Source: Adapted from Rejeski (2003).
[End of figure]
Currently, high-clockspeed technology-based trends alternatively
present an opportunity to reduce highway fatality rates and a threat
that, if not countered, may result in increased fatalities:[Footnote 4]
* Evolving crash avoidance technologies. Key developments include
technologies that are (1) increasingly available in luxury vehicles or
(2) being developed for a future vehicle-road communication system.
They are designed to help drivers by, for example, alerting them to
hazards, potentially decreasing fatality rates.
* Rapidly changing and proliferating electronic driver distractions.
This includes portable devices such as cell phones with text messaging
capabilities that may distract drivers who use them on the road,
risking safety and possibly increasing fatality rates.[Footnote 5]
There is no governmentwide guidance on how federal agencies should
conduct foresight across a time horizon more than 5 years
forward.[Footnote 6] However, when an agency concerned with improving
safety faces a set of potentially significant high-clockspeed
technology-based trends, it may exercise foresight by successfully
carrying out activities such as:
* considering what is known about the safety impact of each high-
clockspeed trend and deciding how to respond to it (for example, by
researching how to shape or counter the trend or taking action),
[Footnote 7]
* reducing uncertainty as needed by developing additional evidence
about the safety impact of such a trend, and:
* communicating with the Congress and others about high-clockspeed and
other trends, agency responses, and policy implications.
We defined these three activities on the basis of literature and
interviews with experts and discuss them further in appendix I. We
recognize that these activities can be challenging because of
uncertainty about the level of safety impact that a new, high-
clockspeed trend may have (that is, its significance) or the kinds of
response options that may be effective; limitations of data systems and
analysis techniques that may be inappropriate for new trends because
they were designed at an earlier time; and the lack of governmentwide
guidance on how agencies should conduct foresight or communicate with
the Congress about new trends and responses. As figure 2 shows, the
three activities may interact in various ways. One example is that new
evidence about a high-clockspeed trend (such as its changing trajectory
over time) may affect agency decisions about whether or how to respond
to that trend.[Footnote 8] Another is that communicating with the
Congress and others may affect an agency's ability to take certain
actions in response to a trend.
Figure 2: Exercising Foresight by Addressing High-Clockspeed Trends:
Three Challenging Agency Activities:
[Refer to PDF for image]
This figure depicts the following information:
Deciding and responding:
Deciding how to respond to or shape each potentially significant high-
clockspeed trend, considering available evidence and uncertainties.
Developing evidence:
Providing additional evidence on the effects of high-clockspeed trends,
to reduce uncertainty (may include devising new data systems or
analysis methods).
Communicating:
Communicating effectively with the Congress and others about high-
clockspeed and other trends, agency responses, and policy implications.
Source: GAO.
[End of figure]
Having discussed your interest in trends that affect future fatality
rates and, in particular, understanding how DOT is addressing high-
clockspeed technology-based trends that may affect these rates, we
focused this report on two such trends--currently evolving electronic
crash avoidance technologies (including in-vehicle technologies and
applications for a future system of road-vehicle communication, as
these relate to passenger vehicles) and rapidly changing electronic
driver distractions--and related challenges that DOT may face.[Footnote
9] Specifically, we examine in this report three foresight issues
corresponding to the challenging activities outlined in figure 2:
1. available evidence on the highway safety impact of two high-
clockspeed trends--that is, the crash avoidance and distraction trends
defined above--and the decisions that DOT, especially the National
Highway Traffic Safety Administration (NHTSA), is making in response;
2. DOT efforts to develop new evidence on the highway safety impacts of
these trends;
3. DOT communications with the Congress and others about these and
other new trends, DOT responses, and policy implications.
To collect data on each of these issues, we obtained documents from and
interviewed officials and staff of relevant DOT administrations, the
Office of the Secretary, and the Transportation Research Board (TRB),
and we interviewed experts from the private sector, universities, and
other organizations--ranging from safety advocates and industry sources
to foresight experts and research methodologists (see app. II). We also
consulted DOT's and TRB's Web sites, attended conferences, and reviewed
varied literature.
To obtain the states' perspectives on available evidence and DOT
decisions, we interviewed state officials, and we reviewed documents
such as newsletters and reports from relevant associations. To further
identify available evidence on evolving crash avoidance technologies
and electronic driver distractions, possible new methods and strategies
for developing evidence, and ways information on trends and
implications for highway safety might be considered and usefully
communicated, we reviewed a variety of literature, including peer
reviewed journal articles, industry reports, and documents from
conference presentations and interviewed experts. We also reviewed a
number of publicly available DOT reports dating from 2001 (or in some
cases earlier) to the present, selected on the basis of their possible
relevance to the future of highway safety. (Appendix III describes our
review of studies of driver phoning and highway safety.[Footnote 10])
Finally, in summarizing evidence on the safety impacts of each high-
clockspeed trend examined here and on DOT responses, we used a
reporting framework that highlights varied levels of evidence and
categories of governance options; this reporting framework is detailed
in the background section of this report.
While several DOT administrations have missions related to highway
safety, we focused our work on issues 1 and 2 on relevant initiatives,
primarily those conducted by NHTSA and the Research and Innovative
Technology Administration (RITA). For issue 3, our review included
NHTSA, RITA, the Federal Motor Carrier Safety Administration (FMCSA),
and the Federal Highway Administration (FHWA), as well as the Office of
the Secretary.
We conducted our performance audit from June 2006 through September
2008 in accordance with generally accepted government auditing
standards. Those standards require that we plan and perform the audit
to obtain sufficient, appropriate evidence to provide a reasonable
basis for our findings and conclusions based on our audit objectives.
We believe that the evidence obtained provides a reasonable basis for
our findings based on our audit objectives.
Results in Brief:
Although high speed technology-based trends are characterized by
limited available data and other uncertainties, the main criteria that
NHTSA officials use in deciding how to respond to such trends do not
address uncertainty. Available evidence confirms that technology-
based, high-clockspeed trends are affecting highway safety and suggests
the potential for stronger impacts in the future. But in addressing
such trends, DOT faces a challenge in making decisions because of the
uncertainties associated with them. For example, DOT lacks strong
quantitative evidence on the specific levels of (1) safety benefits
associated with evolving crash avoidance technologies and (2) safety
risks posed by changing driver distractions--which makes prioritization
difficult. Given available evidence and uncertainties,
* NHTSA has decided to expand its New Car Assessment Program (NCAP)
from evaluating crashworthiness and rollover performance to also
evaluating in-vehicle crash avoidance technologies (to the extent
possible) and providing related consumer information. Additionally,
NHTSA, RITA, and FHWA are working together to explore ways to use crash
avoidance technology in a future vehicle-road information system that
DOT is planning.
* NHTSA is initiating statistical studies to improve evidence on
whether or how electronic driver distractions impact highway safety.
However, NHTSA officials have stated that at this time they will not
initiate action or research aimed specifically at countering the trend
of new driver distractions.
Although there are some uncertainties in both cases, one difference is
that while a large number of crashes are documented as falling into
categories addressed by crash avoidance technologies (for example, rear
end collisions, which are addressed by forward collision warning, or
FCW), less evidence is available that a large number of crashes involve
electronic driver distractions. NHTSA based its decisions on two main
criteria that were met by one trend (crash avoidance) but not the other
(electronic driver distractions): existing evidence that the size of
the problem is large and knowledge of a promising approach to address
it.
However, neither criterion is designed to address the uncertainties
associated with high-clockspeed trends that can affect safety. Some
approaches that might be emphasized to better facilitate decision
making under conditions of uncertainty include, among others,
prioritizing research on ways of responding to trends; weighing
potential gains from focusing on new versus old problems and solutions;
expanding networks; and using anticipatory risk management practices.
DOT also faces challenges in developing additional, higher quality or
more timely evidence on the safety impacts of these trends--and over
time, the changing sizes of these impacts. DOT recognizes that
uncertainty could be reduced by developing better quantitative evidence
on the safety impacts of high-clockspeed trends--including data that
would (1) allow better evaluations of new safety technologies' benefits
and (2) increase information on the level of threat presented by driver
distractions. But DOT faces an evidence-development challenge because,
for high-clockspeed trends, strong data would need to be not only
technically adequate but also timely and able to measure change over
time. Recognizing this, DOT has tried both long-standing and newer
methods. For example, NHTSA's assessment of electronic stability
control analyzed drivers' real-world experiences, including the numbers
of crashes for similar vehicles with and without ESC, based on existing
crash datasets--but this required years of accumulated data.
To achieve a more timely evaluation of new crash avoidance technologies
that are appearing at a relatively fast pace, NHTSA used field
operational tests (FOT) in which subjects drove special vehicles with
sensors and recording equipment installed to provide specific data on
driving experiences. The FOTs tracked each driver for a number of
weeks. However, with relatively short test periods, it is not possible
to compare actual crashes occurring with and without a new technology,
and the resulting data are limited. Additionally, the expense of an FOT
is a limiting factor in terms of repeating tests as new or improved
technologies become available. Similar difficulties face DOT in
tracking new driver distractions and their impact on safety. For
example, existing data cannot capture the extent to which distraction
is a factor in crashes. For the future, new technology might provide
possible directions for collecting additional data more quickly,
allowing frequent updates and reducing uncertainty about high-
clockspeed trends that impact highway safety.
A final challenge for DOT is to effectively inform stakeholders,
including the Congress, about the implications of high-clockspeed
technology trends. Potential avenues for such information might include
DOT's recent framework for reauthorization and planning and
accountability materials. These, however, are not designed or intended
to provide a long-term, comprehensive view and analysis of trends that
could affect highway safety in the future. As a result, while such
materials acknowledge the existence of demographic and other trends,
they do not provide a detailed discussion of the interactions and
implications of these and other trends for the years ahead.
Other materials specifically focused on highway safety, such as the
recent DOT 1.0 Fatality Rate Goal report, which focuses on the year
2011 as requested by congressional committees, are also not designed to
discuss multiple trends and their interactions in time periods such as
2020 and beyond. And they are not issued on a continuing or periodic
basis, since they are produced in response to specific requests. In
addition, these materials are not comprehensive in that some trends are
omitted; for example, they do not include coverage of electronic driver
distractions. While efforts to look toward 2020 and beyond have been
conducted by DOT administrations, and in some cases, an ongoing process
has been set up to conduct futures-related activities, the results of
this work have not been synthesized for use by the Congress. Such work
includes an FMCSA assessment of the operational implications of new
trends for its mission, given motor carrier industry trends for 2025,
and FHWA's review of safety best practices abroad as part of its
ongoing technology scanning program. The production of new futures-
oriented information and synthesis of existing information within DOT
could enhance decision making and deliberations by DOT and
congressional policymakers concerned with how to best position
transportation policies and programs for the future. A review of best
practices and foresight models from the United States and abroad could
provide strategies for DOT's use in dealing with this challenge of
communicating on new trends.
To help ensure that DOT has the most robust information possible when
evaluating policy options in response to new trends in highway
fatalities, even under conditions of uncertainty, and communicating the
most relevant information to the Congress, we are making
recommendations to the Department of Transportation to (1) develop an
approach to guide decision making on high-clockspeed trends that affect
highway safety, (2) evaluate whether or not new data and analytical
systems are needed to better track new trends related to highway
safety, and (3) use a systematic and periodic approach to reporting to
the Congress on a range of trends related to highway safety, including
information on high-clockspeed, technology-based trends.
DOT commented on a draft of this report. DOT disagreed with the first
of these recommendations, giving reasons including difficulty in
foreseeing specific future developments and the significant level of
resources, data, and analysis needed to implement it. We retain this
recommendation because, for example, the techniques we discuss
emphasize flexibility in decision-making strategies to cover a variety
of potential trends, and DOT has recently used such techniques despite
the challenges they cite. DOT did not comment on our other two
recommendations.
Background:
Highway Fatalities:
In 1980 through 1992, the overall highway fatality rate (fatalities per
100 million VMT) declined substantially (as figure 3 shows).[Footnote
11] Progress reducing the rate has since become incremental, and the
future is uncertain.
Figure 3: Actual Overall Fatality Rates and Number of Fatalities, 1975
through 2006, and Rates Projected through 2016:
[Refer to PDF for image]
This figure is a combination vertical bar and line graph depicting the
following data:
Year: 1975;
Number of fatalities: 45,500;
Fatality rate (per 100 million vehicle miles traveled): 3.43.
Year: 1976;
Number of fatalities: 45,523;
Fatality rate (per 100 million vehicle miles traveled): 3.25.
Year: 1977;
Number of fatalities: 47,878;
Fatality rate (per 100 million vehicle miles traveled): 3.26.
Year: 1978;
Number of fatalities: 50,331;
Fatality rate (per 100 million vehicle miles traveled): 3.26.
Year: 1979;
Number of fatalities: 51,093;
Fatality rate (per 100 million vehicle miles traveled): 3.34.
Year: 1980;
Number of fatalities: 51,091;
Fatality rate (per 100 million vehicle miles traveled): 3.35.
Year: 1981;
Number of fatalities: 49,301;
Fatality rate (per 100 million vehicle miles traveled): 3.17.
Year: 1982;
Number of fatalities: 43,945;
Fatality rate (per 100 million vehicle miles traveled): 2.76.
Year: 1983;
Number of fatalities: 42,589;
Fatality rate (per 100 million vehicle miles traveled): 2.58.
Year: 1984;
Number of fatalities: 44,257;
Fatality rate (per 100 million vehicle miles traveled): 2.57.
Year: 1985;
Number of fatalities: 43,825;
Fatality rate (per 100 million vehicle miles traveled): 2.47.
Year: 1986;
Number of fatalities: 46,087;
Fatality rate (per 100 million vehicle miles traveled): 2.51.
Year: 1987;
Number of fatalities: 46,390;
Fatality rate (per 100 million vehicle miles traveled): 2.41.
Year: 1988;
Number of fatalities: 47,087;
Fatality rate (per 100 million vehicle miles traveled): 2.32.
Year: 1989;
Number of fatalities: 45,582;
Fatality rate (per 100 million vehicle miles traveled): 2.17.
Year: 1990;
Number of fatalities: 44,599;
Fatality rate (per 100 million vehicle miles traveled): 2.08.
Year: 1991;
Number of fatalities: 41,508;
Fatality rate (per 100 million vehicle miles traveled): 1.91.
Year: 1992;
Number of fatalities: 39,230;
Fatality rate (per 100 million vehicle miles traveled): 1.75.
Year: 1993;
Number of fatalities: 40,134;
Fatality rate (per 100 million vehicle miles traveled): 1.75.
Year: 1994;
Number of fatalities: 40,718;
Fatality rate (per 100 million vehicle miles traveled): 1.73.
Year: 1995;
Number of fatalities: 41,817;
Fatality rate (per 100 million vehicle miles traveled): 1.73.
Year: 1996;
Number of fatalities: 42,065;
Fatality rate (per 100 million vehicle miles traveled): 1.68.
Year: 1997;
Number of fatalities: 42,013;
Fatality rate (per 100 million vehicle miles traveled): 1.64.
Year: 1998;
Number of fatalities: 41,501;
Fatality rate (per 100 million vehicle miles traveled): 1.58.
Year: 1999;
Number of fatalities: 41,717;
Fatality rate (per 100 million vehicle miles traveled): 1.55.
Year: 2000;
Number of fatalities: 41,945;
Fatality rate (per 100 million vehicle miles traveled): 1.53.
Year: 2001;
Number of fatalities: 42,116;
Fatality rate (per 100 million vehicle miles traveled): 1.51.
Year: 2002;
Number of fatalities: 43,005;
Fatality rate (per 100 million vehicle miles traveled): 1.51.
Year: 2003;
Number of fatalities: 42,884;
Fatality rate (per 100 million vehicle miles traveled): 1.48.
Year: 2004;
Number of fatalities: 42,836;
Fatality rate (per 100 million vehicle miles traveled): 1.45.
Year: 2005;
Number of fatalities: 43,443;
Fatality rate (per 100 million vehicle miles traveled): 1.45.
Year: 2006;
Number of fatalities: 42,642;
Fatality rate (per 100 million vehicle miles traveled): 1.42.
Year: 2008;
Assumes that the fatality rate for 2006 is maintained through 2016:
1.42.
Assumes that the fatality rate declines to 1.2 by 2016: 1.38.
Year: 2010;
Assumes that the fatality rate for 2006 is maintained through 2016:
1.42.
Assumes that the fatality rate declines to 1.2 by 2016: 1.33.
Year: 2012;
Assumes that the fatality rate for 2006 is maintained through 2016:
1.42.
Assumes that the fatality rate declines to 1.2 by 2016: 1.29.
Year: 2014;
Assumes that the fatality rate for 2006 is maintained through 2016:
1.42.
Assumes that the fatality rate declines to 1.2 by 2016: 1.24.
Year: 2016;
Assumes that the fatality rate for 2006 is maintained through 2016:
1.42.
Assumes that the fatality rate declines to 1.2 by 2016: 1.2.
Source: GAO analysis of National Highway Traffic Safety Administration
(NHTSA) data.
[End of figure]
The declines in the fatality rate between 1980 and 1992 were associated
with (1) the development and spread of technologies that improve
crashworthiness such as seatbelts, airbags, and improved car design--as
well as a related behavior change--increased seatbelt use; (2) declines
in driver alcohol use; and (3) according to DOT, other factors, such as
roadway improvement. DOT estimates that technology-related improvements
saved over 325,000 lives on U.S. roads from 1960 to 2002 (Kahane 2004).
But despite continuing improvement in seatbelt use, driver alcohol use
has plateaued, and highway fatalities remain the leading cause of death
for children, teens, and young adults (NHTSA 2008c; GAO 2008c). While
DOT's prior goal for 2008 was an overall fatality rate of 1 fatality
per 100 million vehicle miles traveled, the figure is still close to
1.4 fatalities per 100 million VMT.[Footnote 12] The dashed and dotted
lines of figure 3 are based on the alternative assumptions about future
fatality rates that a DOT official used to project fatalities for 2016
(Kanianthra 2007).[Footnote 13]
The annual number of fatalities (see grey bars of figure 3) is related
to the fatality rate and VMT--both of which are uncertain in the
future.[Footnote 14] VMT reflects population size and mobility
patterns, which in turn reflect factors such as the economy, fuel
costs, and the availability of alternative modes of transportation. If
trajectories such as those projected for the fatality rate in figure 3
were to extend through 2020--and if VMT were to remain steady--highway
fatalities over the next 12 years could be close to 500,000.[Footnote
15] Declining VMT would be likely to be associated with reduced
fatalities, but efforts to reduce future fatality rates would remain a
highly important priority for the federal and state governments.
Whether the overall fatality rate will increase or decline between now
and 2020 is uncertain. Fatalities may rise because of anticipated
increases in some vulnerable road-user groups--including older drivers,
occupants of small cars (a group that may increase if small-car sales
continue to surge in response to rising fuel prices), and
motorcyclists.[Footnote 16] Conversely, fatalities may decline because
of new technology or other ongoing safety efforts (such as programs to
make the road environment safer through better materials or signs or to
emphasize speeding enforcement). However, how DOT deals with fast-paced
technological changes may also play a role in the trajectory of the
highway fatality rate. While many new technology developments may
affect the future of highway safety, experts view two--both of them
high-clockspeed trends--as clearly significant. Post-ESC crash
avoidance technologies present an apparent opportunity for improving
highway safety--ESC, in particular, is likely to provide considerable
safety benefits (see Dang 2007)--and electronic driver distractions
present an apparent threat to safety:
Trends Related to Post-ESC Crash Avoidance Systems:
Several factors are contributing to the nature and clockspeed of the
crash avoidance trend:
* Automobile manufacturers are developing new kinds of new crash
avoidance systems (electronic systems aimed at assisting drivers),
making them available in luxury cars and increasingly including them in
other new models.[Footnote 17]
* Statements by recent NHTSA administrators and others point to a
recognition that automobile manufacturers are developing new safety
technologies at an unprecedented rate.
* Rapid advances in sensing and microprocessor technologies have
allowed a faster pace for the introduction of in-vehicle safety
systems, according to an industry representative.
Post-ESC in-vehicle crash avoidance technologies include systems to
warn drivers who are drifting out of their lane or risking rear-end
collisions. Several in-vehicle systems were developed by automotive
suppliers and introduced by automobile manufacturers within about 8
years. Figure 4 lists a number of these systems.[Footnote 18]
Figure 4: Evolving In-Vehicle Crash Avoidance Technologies:
[Refer to PDF for image]
This figure is an illustration of when evolving in-vehicle crash
avoidance technologies became available. A timeline is shown indicating
1995 through 2008. The following technologies became available during
that time span:
Electronic stability control;
Night vision;
Adaptive cruise control;
Roll stability control;
Lane departure warning;
Blind spot warning;
Brake assist;
Forward collision warning;
Automatic braking.
Source: Adapted from information provided by NHTSA.
Note: The in-vehicle technologies in this figure were all introduced
within the period shown. However, different automobile manufacturers
introduced different technologies at different times, and the specific
order of introduction may have varied.
[End of figure]
The crash avoidance technologies listed in figure 4 are already
available in some new cars, especially luxury cars, and the percentage
of cars on the road that have these technologies is likely to increase.
But despite the relatively fast pace at which these technologies are
introduced, increases in the percentage of cars with these technologies
will be relatively slow. That is, because the average car now stays on
the road for several years, it will be some time before new cars with
the new technologies are sold to enough customers to become a high
percentage of vehicles on the road.[Footnote 19]
Automobile manufacturers, DOT, and others are planning other approaches
to crash avoidance, including the following:
* Vehicle-to-vehicle communications (V2V). Some automobile
manufacturers are designing V2V communications so that, for example, a
car encountering an icy patch could signal this potential hazard to
following cars.
* Safety applications planned for Vehicle Infrastructure Integration
(VII). DOT, in partnership with others, is planning a "smart roads" VII
system of real time vehicle-road communication. Applications could
include crash avoidance--as well as electronic payments for tolls,
parking, and fuel and other applications.[Footnote 20]
Experts told us that V2V and VII (that is, vehicle-to-vehicle
communications and vehicle infrastructure integration) are intended to
provide each car with information beyond what its own sensors can
detect--for example, information about a road blockage (such as a
recent crash) or other hazard around a curve or over a hill. Figure 5
illustrates the three interrelated types of crash avoidance
applications.
Figure 5: Three Major Types of Crash Avoidance Applications:
[Refer to PDF for image]
This figure contains illustrations of three types of crash avoidance
applications as follows:
New in-vehicle crash-avoidance technologies such as Lane Departure
Warning (LDW), which warns drivers to help prevent drifting into
another lane.
Vehicle-to-vehicle (V2V) communication: One vehicle can communicate a
warning to others; for example, a lead vehicle might warn following
vehicles about icy patches on the roadway ahead.
Crash avoidance applications for the planned Vehicle Infrastructure
Integration (VII) initiative, in which vehicles would communicate with
roadside devices.
Source: GAO.
[End of figure]
Some crash avoidance applications use audible or tactile warnings.
[Footnote 21] Others provide visual information on screens-- for
example, backup cameras, night vision screens, and V2V signal screens.
[Footnote 22] Issues for safety technologies such as these can be
complex, depending on how drivers interface with or react to new
signals or systems (see Wochinger and others 2008). Even more basic
technologies or improvements that provide a safer vehicle or contribute
to a safer road environment (for example, widening roads) may encourage
some to drive faster or less carefully, reducing the overall safety
benefit below what had been anticipated.[Footnote 23]
Automobile manufacturers, suppliers, and various experts view crash
avoidance technologies as the wave of the future because of their
potential to reduce the number and severity of accidents. Studies of
ESC promise eventual fatal crash reductions by nearly one-third, and in
2007 the chairman of the National Transportation Safety Board (NTSB)
stated that:
"while we accomplished much ... to improve the crashworthiness of
automobiles, we have reached some practical limits in combating the
physical forces involved in crashes. It is time to ... enter a new era
where technology will help us prevent accidents." (Rosenker 2007):
Others, including a recent NHTSA administrator, have expressed similar
views. A NHTSA official told us that some new in-vehicle crash
avoidance technologies, such as lane departure warning (LDW), would be
likely to mitigate the effect of electronic driver distractions, such
as driver cell phone use. Effective crash avoidance and other
technologies might also help counter the effects of driving while
drowsy or with elevated blood alcohol content, and even failing to wear
seat belts.
Trends Related to Evolving Electronic Driver Distractions:
Another major technology-based trend relates to the growing use of
electronic devices that could distract drivers. The electronics
industry is developing new devices or new features for existing devices
for portable communication, information access, and entertainment and
introducing them to consumers, who purchase and use them at an ever
accelerating rate. Using electronic devices is distinct from other
behavior that can distract drivers, such as eating or grooming, because
the new devices are proliferating and evolving quickly with great
complexity (see figure 6).
Figure 6: The Evolution and Spread of Portable Electronic Devices in
the United States:
[Refer to PDF for image]
This figure illustrates a timeline indicating the evolution and spread
of portable electronic devices in the United States, as follows:
1994:
Cell phone subscriptions number 24 million.
1999:
Personal digital assistant (PDA) becomes widely available;
2000:
Cell phone subscriptions increase to 109 million;
Text messages number about 14 million per month.
2002:
Advanced hands-free cell phones (small earpiece without wire) become
widely available.
2004:
Cell phones with cameras become widespread;
Text messages increase to 4.7 billion per month;
Consumer cell phones with GPS functions become widely available.
2006:
Web accessed via cell phones by 20 million subscribers;
Text messages increase to 18.7 billion per month.
2007: Cell phone subscriptions increase to 255 million;
MP3 player kits for connection to car audio system reach 14% of U.S.
households.
Source: GAO analysis based on expert opinion and data from Consumer
Electronics Association (CEA), CTIA—The Wireless Association®,
Telecommunications Industry Association (TIA), and documents such as
industry reports and estimates.
[End of figure]
Indicators of rapid change include:
* the rate at which the average cell phone subscriber replaces a hand
set--about every 17 months (which allows many consumers to obtain new
features quickly);
* the number of minutes of cellular usage, which has increased at an
exponential rate (see figure 7);
* new phones that include streaming TV and video calling and MP3
players that feature screens for movies; and:
* new kinds of screens (see figure 8).
Figure 8: Overall Minutes of Cellular Use in the United States, 1997-
2007:
[Refer to PDF for image]
This figure is a line graph depicting the following data:
Year: 1997;
Minutes of use (billions): 63.
Year: 1998;
Minutes of use (billions): 89.
Year: 1999;
Minutes of use (billions): 148.
Year: 2000;
Minutes of use (billions): 259.
Year: 2001;
Minutes of use (billions): 457.
Year: 2002;
Minutes of use (billions): 620.
Year: 2003;
Minutes of use (billions): 830.
Year: 2004;
Minutes of use (billions): 1101.
Year: 2005;
Minutes of use (billions): 1495.
Year: 2006;
Minutes of use (billions): 1798.
Year: 2007;
Minutes of use (billions): 2119.
Source: GAO presentation of data provided by CTIA—The Wireless
Association® (used by permission).
[End of figure]
Figure 8: Touch Screens and Virtual Screens:
[Refer to PDF for image]
This figure contains the following three illustrations of touch screens
and virtual screens:
Example of flowing touch screen;
Example of clear "video eyeglasses";
Example of virtual big screen with see-around opaque eyewear (iPod plus
myvu eyewear = big screen viewing).
Source: GAO (touch screen), eyeglasses image courtesy of Lumus Ltd.,
and Myvu Corporation (virtual big screen).
[End of figure]
DOT Units with Highway Safety Missions:
Several DOT administrations are involved in monitoring and addressing
trends that affect highway safety. NHTSA has historically been charged
with setting new car safety standards and testing new cars for
compliance with those standards. It is also responsible for conducting
research and administering grant programs designed to assist states in
addressing driver behavior issues, such as seatbelt and booster seat
use, and state alcohol-impaired driving programs. NHTSA has addressed
new car safety through research, regulation, and consumer information.
For example, NHTSA's evidence-based New Car Assessment Program has been
designed to increase consumer demand for safer cars and thus encourage
automobile manufacturers to improve safety.
Working with the states, NHTSA has also addressed driver behavior
issues by (1) awarding and overseeing grants to aid states in
developing safety programs and enforcing highway safety laws; (2)
conducting public information campaigns on the effectiveness of
countermeasures, such as seatbelts; and (3) promoting compliance with
legislative requirements that make grants contingent on state action,
such as passing a child booster seat law.[Footnote 24]
NHTSA also houses the National Center for Statistics and Analysis
(NCSA), which maintains crash datasets--such as the Fatality Analysis
Reporting System (FARS), which provides data on all fatal motor vehicle
crashes on U.S. roads--and other datasets based on observational and
telephone surveys.[Footnote 25]
RITA, another DOT administration, coordinates the department's research
programs. For example, RITA is the lead administration for the VII
project through its Intelligent Transportation Systems program, which
directs research investments in developing future intelligent
transportation systems. RITA's Administrator is also principal science
adviser to the Secretary on science and technology matters and oversees
DOT's Volpe National Transportation Systems Center and more than 60
University Transportation Centers.
While TRB is not a part of DOT, it is a part of the National Academies
and conducts a variety of activities. Its oversight of research
includes safety research, often in coordination with DOT.[Footnote 26]
The responsibilities of the Office of the Secretary of Transportation,
FMCSA, and FHWA also relate to highway safety. For example, the Office
of the Secretary of Transportation provides policy development,
oversight, and coordination for the overall planning and direction of
DOT, including DOT's strategic plan and budget request. DOT's strategic
planning efforts are guided by governmentwide requirements. Although
some foresight-related activities are conducted by the Office of the
Secretary, there is no specific foresight unit. The Office of the
Secretary also performs or sponsors some research independently of
other DOT agencies.
FMCSA is charged with reducing crashes involving commercial vehicles
(large trucks and buses), including developing and enforcing related
regulations. FHWA also performs safety-related functions. Notably, FHWA
administers the Highway Safety Improvement Program, through which it
allocates money to states annually for infrastructure-related safety
improvements, and the State and Community Highway Safety Grant Program,
which supports state highway safety programs designed to reduce traffic
crashes and resulting deaths, injuries, and property damage. FHWA also
adopts standards for the design of roadways receiving federal funds and
told us that "safety is fully considered when developing these
standards." Finally, FHWA has a research mission; its current work
includes creating a dataset of roadway characteristics identifiable
with global positioning system (GPS) coordinates. In addition, FHWA has
initiated the Model Minimum Inventory of Road Elements (MMIRE) project,
designed to set a baseline of information for states to collect and use
in selecting safety-related roadway improvements.
Foresight Reporting Framework for High-Clockspeed Trends:
To help describe how DOT is considering evidence and otherwise
addressing post-ESC crash avoidance technologies and electronic driver
distractions, we developed a framework with categories for levels of
evidence and governance options. Specifically, this framework provides
a way to summarize and analyze:
* the level of available evidence on each trend's impact and:
* the various governance options that DOT has employed in response or
could potentially employ.
Evidence about an evolving or fast-changing trend's safety impact may
vary across a continuum of certainty or level of knowledge. Toward one
end of the continuum, relatively weak early signals may suggest that a
new trend or phase of a trend is likely and might impact highway
safety, now or in the future. Early signals may consist, for example,
of anecdotal evidence, such as a few widely reported crashes involving
driver use of a new technology or interviews with key industries in
which possible plans for future technologies are outlined. Toward the
opposite end of the continuum, strong quantitative data would, if
available, provide evidence of the scope and magnitude of the impact on
highway safety associated with an evolving trend--as well as, where
relevant, how safety impacts have changed over time or are currently
changing.[Footnote 27] More than one level of evidence may be relevant
to a high-clockspeed evolving trend, because different sets of evidence
may apply to various phases of such a trend. Figure 9 shows three
levels of evidence on safety impacts.
Figure 9: Foresight Reporting Framework for High-Clockspeed Trends:
[Refer to PDF for image]
This figure contains the following information:
Levels of evidence on safety impacts of a new technology-based
opportunity or threat:
Level 1: Early signals:
Networking with industry to learn about products "in the pipeline,"
scanning for instances or qualitative data indicating possible impacts
on safety. May signal future developments or interrelationships among
trends
Level 2: Confirming qualitative or limited quantitative evidence:
Results from studies or tests that confirm the existence of an evolving
opportunity for or threat to safety; suggest a general level of impact
on safety or how that impact is changing[A].
Level 3: Strong quantitative evidence:
Results from studies that quantify the magnitude of the overall safety
impact of an opportunity or threat; may also define the pace of change
and anticipated trajectory of an evolving trend[A].
Governance options for addressing a new safety opportunity or threat:
Option A: No self-initiated response:
* Decision not to self-initiate efforts to pursue opportunities or
counter threats at this time.
Option B: “Starting point” actions:
* Discussion forums (issue clarification, agenda setting), early work
to develop policy proposals or priorities, outlining a "vision."
Option C: Research to explore or evaluate approaches to action or to
stimulate action; early interventions:
* Research to explore new countermeasures or other programs; evaluation
and demonstrations programs;
* Information programs designed to be supportable by Level 1 or Level 2
evidence;
* Involve manufacturers or suppliers in conducting research to
encourage development of certain types of new safety products.
Option D: Mid-level interventions:
* Consumer information programs that include very specific or
comparative safety information or require competing products to be
directly compared in terms set by the government;
* Provide information to states; encourage states to develop new
programs or pass new laws.
Option E: Stronger interventions:
* Regulation;
* Federal incentives;
* Grants to states.
Source: GAO.
Note: "Governance options" refers to tools or strategies for research
and action that agencies can use to pursue goals; they include official
tools such as regulation and informal approaches such as networking.
Various agencies may use different terminology for this and other
concepts in this figure. The categories shown illustrate issues
relevant to this report and may not be comprehensive; additionally,
more than one category may be relevant to some DOT responses.
[A] Level 2 evidence falls short of strong quantitative evidence,
defined as valid, reliable, and generalizable (Level 3). To illustrate,
one aspect of validity for evidence of safety impact concerns whether
data include information on serious or fatal crashes. (See appendix III
for a further definition of validity, reliability, and
generalizability.) Other key characteristics relevant to strong
evidence on high-clockspeed trends include timeliness and measurement
of change. Results from multiple quantitative studies conducted by
different investigators are another indication of strength.
[End of figure]
Describing levels of available evidence on safety impacts can be
important, because stronger evidence may allow an agency to more
confidently make decisions about whether or how to respond to a trend.
At the same time, early signals about possible safety impacts, now or
in future, may suggest anticipatory research or action.
DOT's responses to a technology-based trend (that is, DOT choices of
governance options) may be based on its own judgments about evidence as
well as the requirements or suggestions of others--for example,
statutory mandates. Options available to DOT include (1) forgoing the
"self-initiation" of research or action to promote or counter a
technology-based trend; (2) pursuing exploratory research and
development efforts to identify what kinds of actions might effectively
address a new trend or sponsoring limited demonstration programs to
facilitate broader action in the future; (3) providing the general
public with consumer information or conducting an information campaign;
(4) providing information to states (an option that may be especially
relevant for driver behaviors that the states regulate); and (5)
regulation on its own initiative or as mandated by the Congress. Figure
9 illustrates these governance options.
Agencies may pursue various options through traditional policy tools or
organizational networking or a combination of these.[Footnote 28] The
various options are not mutually exclusive, and a department such as
DOT could respond to a single trend by employing multiple options--
self-initiating some, carrying out others in response to congressional
interest or directives, and so forth.[Footnote 29] Additionally, some
options may be pursued in preparation for using others--for example,
research to explore or assess possible interventions may be pursued in
preparation for later action.
Finally, there is not a prescribed one-to-one link between a particular
level of evidence on safety impact and a corresponding governance
option. This is because in choosing an option, an agency may weigh many
interrelated factors and trade-offs. More generally, proposed safety
improvements may sometimes be weighed against mobility needs, possible
impacts on congestion, or other issues.
Uncertainty about New Opportunities and New Threats from High-
Clockspeed Trends Challenges DOT Decision Makers:
High-clockspeed trends based on fast-paced innovation and consumer
adoption of new products can be associated with high levels of
uncertainty, and making decisions about when and how to address such
trends presents a fundamental foresight challenge (Rejeski 2003, 56).
(See fig. 10.)
Figure 10: Exercising Foresight by Addressing High-Clockspeed Trends,
with Deciding and Responding Highlighted:
[Refer to PDF for image]
This figure depicts the following information, with the first
highlighted:
[Highlighted: Deciding and responding:
Deciding how to respond to or shape each potentially significant high-
clockspeed trend, considering available evidence and uncertainties].
Developing evidence:
Providing additional evidence on the effects of high-clockspeed trends,
to reduce uncertainty (may include devising new data systems or
analysis methods).
Communicating:
Communicating effectively with the Congress and others about high-
clockspeed and other trends, agency responses, and policy implications.
Source: GAO.
[End of figure]
Evidence on safety impacts shows that evolving post-ESC crash avoidance
technologies present an opportunity to enhance safety and that
electronic driver distractions pose a threat to safety. But there are
uncertainties about the magnitude of safety impacts and the
effectiveness of possible actions. As a result, it is uncertain how
many lives might be saved by pursuing the crash-avoidance opportunity
or countering the distraction threat. In the absence of a
governmentwide structure for exercising foresight across a horizon more
than 5 years forward--and in the face of potentially significant,
although somewhat uncertain, high-clockspeed trends that can affect
highway safety (with one trend presenting a new opportunity, the other
a threat)--DOT administrations are responding by actively attempting to
shape one trend (crash avoidance) but have decided not to counter the
other (electronic distractions) at this time. The main criteria NHTSA
officials told us they used in making these decisions are that (1)
existing quantitative evidence demonstrates that a sizable safety
problem already exists (for example, roughly as many fatalities are
counted for the new problem as for a recognized major problem such as
lack of seatbelt use) and (2) a promising strategy exists for dealing
with the new problem. These criteria do not specifically target the
uncertainty of high-clockspeed trends, which may be greater than for
slower trends or long-standing problems. Literature and experts suggest
a variety of approaches to dealing with uncertainty in decision making,
including anticipatory risk management, prioritization of research on
new strategies, and expansion of networking, among others.
Although DOT Administrations Face Some Uncertainty Regarding Crash
Avoidance Technologies, They Are Moving to Actively Shape This Trend:
Some evidence exists regarding the effect of crash avoidance
technology. NHTSA has obtained early evidence on new safety
technologies that automobile manufacturers are developing by attending
transportation or safety conferences and by networking with automobile
manufacturers and suppliers about plans and early tests. Evidence from
such activities corresponds to Level 1 in figure 11.
Figure 11: Foresight Reporting Framework: Post-ESC In-Vehicle Crash
Avoidance Technologies:
[Refer to PDF for image]
This figure contains the following information:
Evidence of safety benefits for post-ESC in-vehicle crash avoidance
technologies:
Level 1: Early signals:
* Information from automobile manufacturers and suppliers on early
tests and plans for new technologies (available to NHTSA). (Available
evidence, main response to Option D)
Level 2: Confirming qualitative or limited quantitative evidence:
Results from small-sample field operational studies (FOTs) indicate two
technologies (LDW and FCW) have some potential benefit. FOT results
combined with data on the frequency of relevant precrash scenarios
suggest overall levels of safety effectiveness. (Available evidence,
main response to Option D)
Level 3: Strong quantitative evidence:
No study has thus far demonstrated the extent to which any post-ESC in-
vehicle crash avoidance technology will prevent actual crashes.
NHTSA responses to post-ESC in-vehicle technologies including action or
research to promote safety benefits:
Option A: No self-initiated response:
* NHTSA did not select this option; see below.
Option B: “Starting point” actions:
* NHTSA has moved beyond starting point actions through pursuing the
activities shown in Option C.
Option C: Research to explore or evaluate approaches to action or to
stimulate action; early inventions:
* NHTSA has conducted “action research” aimed at stimulating
development of various crash avoidance technologies[A]; (Selected
response);
Option D: Mid-level interventions:
* NHTSA's main response is to require that, as part of its new NCAP
program, new cars sold in the U.S. will include sticker information on
whether LDW and FCW systems are standard or optional[B]; (Selected
response);
Option E: Stronger interventions:
* NHTSA has opted not to pursue regulation for new crash avoidance
technologies other than ESC.
Source: GAO.
[A] A NHTSA official told us that its Advanced Crash Avoidance
Technologies (ACAT) program, in which NHTSA partners with automobile
manufacturers to improve information on safety impacts of crash
avoidance technologies, is intended to encourage manufacturer efforts
to develop such technologies. Also, NHTSA has conducted varied
fundamental research aimed at understanding crash processes or factors
relevant to crash avoidance technologies--for example, human factors
research. This could help support judgments about precrash scenarios
and new technologies that might be successful in addressing them. A
possible alternative or supplementary option would consist of other
forms of consumer education on new technologies, such as how to best
use them.
[B] 73 Fed. Reg. 40,016 (July 11, 2008).
[End of figure]
NHTSA has obtained systematic evidence on safety effectiveness from
FOTs on two technologies that have been shown to have some safety
benefit--LDW and FCW systems.[Footnote 30] The NHTSA-sponsored FOTs are
not large or lengthy enough to provide definitive quantitative evidence
on levels of safety benefit or to compare serious or fatal crashes in
vehicles with or without each new technology (as we discussed later).
However, NHTSA has combined the FOT data with other information on the
frequency of relevant crash and precrash scenarios to obtain estimates
of "safety effectiveness."[Footnote 31] For example, rear-end
collisions, which are addressed by FCW, are a frequent type of crash.
These combined estimates represent "confirming qualitative or limited
quantitative evidence"--Level 2 evidence in figure 11. However, there
is a lack of strong, quantitative (Level 3) evidence for safety
benefits such as how many serious crashes would be prevented by LDW and
FCW.
In response to the existing evidence on safety impacts, NHTSA has made
decisions to actively shape the in-vehicle crash avoidance trend by
pursuing two governance options:
* The first is to spur industry safety innovations and encourage
manufacturers and suppliers to develop such technologies by involving
them in related research. Using existing resources, NHTSA has carried
out some initiatives to achieve this goal (Option C in figure 11).
* The second, NHTSA's main response to this trend thus far, is to
capitalize on the potential opportunities afforded by crash avoidance
technologies. As of model year 2010, NHTSA will add information on LDW
and FCW (and ESC) systems to NCAP labels on new cars in order to
indicate whether such technologies are standard or optional (Option D
in Figure 11). This is aimed at accelerating consumer interest in
buying cars with new safety technologies.[Footnote 32]
NHTSA officials told us that in addition to considering the FOT results
on LDW and FCW and other relevant results, their decision to encourage
these in-vehicle technologies was inspired by the demonstrated success
of ESC, which is expected to cut fatalities by a substantial margin
once widely deployed, and their knowledge of the variety of crash
avoidance technologies being designed by suppliers and manufacturers.
They also said that congressional interest in crash avoidance, as well
as our recommendations concerning the need to update NCAP, influenced
their decision (GAO 2005b, 58-59). Additionally, NCAP has successfully
encouraged crashworthiness in vehicles, and NHTSA decision makers, as
well as automobile manufacturers, anticipate that expanding NCAP to
include crash avoidance technologies will have a positive result
(although this has not been tested).[Footnote 33]
Maximizing the benefits of evolving crash avoidance technologies may be
challenging, however, given various uncertainties about safety impacts,
including the limitations of data and interactions between crash
avoidance and other trends. For example, an article by NHTSA staff
raised a variety of issues about the future of in-vehicle crash
avoidance as the population of older road users increases--ranging from
the potential for crash avoidance technologies to substantially enhance
both safety and mobility for this vulnerable group to the need to
design technologies with this group's limitations in mind (Band and
Perel 2007). However, it was noted that crash avoidance technologies
"might encourage older adults to continue driving well beyond when they
would ordinarily cease operating vehicles," thus raising risks (based
on increased exposure to crash risks because of more vehicle miles
traveled). The implication was that such risks would not be raised so
high as to negate overall safety benefits, but the article did not
analyze relative risks. In addition, representatives of the automobile
industry have said that consumer training in the use of new
technologies could be key to maximizing safety benefits. The final
notice of changes to NCAP acknowledges this point and states that NHTSA
will continue to conduct marketing studies in order to assess consumer
reaction to new technologies.
For the longer term, DOT faces uncertainties about how the crash
avoidance trend will impact safety in the future, depending on whether
or how it moves toward V2V and VII. Some manufacturers are working to
develop V2V, but it is unclear whether this will be widely adopted or
how interoperability will be achieved. NHTSA is working with FHWA and
RITA to develop possible crash avoidance applications for VII.[Footnote
34] Although some DOT administrations and others have long discussed
the potential safety benefits of VII and RITA's Volpe Center is
preparing a benefit-cost analysis, actual data on VII's safety or other
benefits are limited, and how government or private sector groups would
fund VII is unclear. RITA envisions a public-private partnership that
will implement a high-speed data network for collecting anonymous data
from moving vehicles and providing information to them.[Footnote 35] As
figure 12 shows, the VII system would:
* provide a link to each vehicle equipped with an onboard transceiver
by roadside devices or other wireless technologies, using multiple
paths of communications to ensure redundancy[Footnote 36] and:
* accomplish crash avoidance by, for example, alerting a driver to
hazards, including some that neither the driver nor sensors in the
vehicle would otherwise detect in time to stop (such as a runaway truck
heading for an intersection obscured by buildings or a crash around a
sharp curve in the roadway ahead).[Footnote 37]
Figure 12: How VII Would Operate:
[Refer to PDF for image]
This figure contains three illustrations depicting the following
information:
1) Vehicles would be equipped with a transceiver called an On Board
Unit (OBU). Roadway infrastructure, including signalized intersections
and highways, would be equipped with transceivers called Roadside Units
(RSU).
2) The OBU would collect GPS data and other sensor data from various
systems in the vehicle and would anonymously send and receive data to
and from RSU, as well as to and from other vehicles as they passed by
on the roadway.
3) Data transmitted from roadside to vehicle could warn a driver about
dangerous road conditions, reducing accidents. Additionally, data
transmitted from vehicle to roadside could anonymously provide road
condition and traffic information to transportation agencies, enabling
improved highway safety and more effective operation of transportation
systems.
In this example a runaway truck (A) transmits a warning signal to a
roadside unit (B), which in turn sends a warning signal to a truck (C)
about to enter the intersection and other vehicles (D) on the roadway.
Source: GAO.
[End of figure]
NHTSA officials stated that some opportunities existed to incorporate
VII into prior crash avoidance initiatives involving communications.
For example, the Cooperative Intersection Collision Avoidance System
(CICAS), which is managed by a team including NHTSA and FHWA staff, and
Effectiveness of Vehicle Safety Communications Applications (EVSCA)
initiatives could utilize VII technology.[Footnote 38] The future
system would be implemented nationwide, in urban and rural areas, and
could host a broad range of potential applications not geared to
safety, ranging from toll payment to information on congestion
reduction to a variety of commercial uses. Some have suggested that
commercial applications of these might, if implemented, produce revenue
to help fund the system.[Footnote 39] DOT officials told us they expect
to decide on whether it is feasible to move forward with field
operational tests of a VII system sometime in fiscal year 2009.
Looking forward, VII faces a number of uncertainties, including limited
data on its effectiveness and unclear sources of funding, but NHTSA and
other DOT administrations, such as those participating in the CICAS
team, are attempting to include a crash avoidance safety component in
VII plans. In terms of the decision-option categories we developed for
this report, DOT efforts to develop VII-related applications fall into
the category of starting point actions (Option B in figure 9), with
testing of safety applications falling into exploratory research
(Option C in figure 9).
Although Evidence on Electronic Driver Distractions Points to Safety
Risks, NHTSA Has Not Initiated Responses, Cites Uncertainties:
Studies confirm that driver phoning raises safety risks. Based on a
combination of NHTSA publications, peer-reviewed articles, and recent
surveys, the overall safety impact of driver phoning appears to be
substantial:
* A recent NHTSA-sponsored literature review concluded that (although
key studies cited were conducted in Australia and Canada rather than
the United States) "the available evidence suggests that cell phone use
increases drivers' crash risk by a factor of 4" (Ranney 2008, iii).
* Our review confirms that driver calls made with portable phones
increase the risk of a crash. We sampled 13 primary studies that
evaluated this issue, including those we identified in peer-reviewed
journals and an additional NHTSA-sponsored study known as the "100 car
study."[Footnote 40] Of the 13 studies, 12 reported an increase in the
risk of an actual crash, simulated crash, or near miss. Ten studies
quantified the increase in risk, and of the ten, seven estimated an
overall doubling or higher increase in risk (see appendix III).
* Two nationwide self-report surveys conducted in 2006 indicate that a
majority of respondents admit to phoning while driving.[Footnote 41]
NHTSA's roadside observations indicate that at any one moment in 2007,
6 percent of drivers were observed holding phones to their ears (NHTSA
2008a).[Footnote 42] NHTSA used data from a questionnaire survey to
adjust for unobserved hands-free use, resulting in an overall phoning
estimate of 11 percent in 2007.[Footnote 43]
These results are "confirming qualitative or limited quantitative
evidence"--Level 2 evidence of safety impact in figure 13. However,
there is a lack of strong quantitative (Level 3) evidence of a major
safety impact.
Figure 13: Foresight Reporting Framework: Evolving Electronic Driver
Distractions:
[Refer to PDF for image]
This figure contains the following information:
Evidence of safety impact of electronic driver distractions[A]:
Level 1: Early signals:
Varied information suggests that:
* driver texting and other new complex uses of technology may degrade
safety more than voice calls;
* driver use of future handheld devices may be riskier;
* the frequency and extent of driver use of existing or new devices (or
both) may increase in the coming years; (Available evidence, part of
main response to Option A);
Level 2: Confirming qualitative or limited quantitative evidence:
Taken together, results from diverse studies—including driving
simulator studies, surveys, case-control studies that match crashes
with checks of driver cell-phone records, the NHTSA-sponsored 100-car
study, and NHTSA roadside observations—confirm that:
* driver calls made with portable devices impact safety negatively;
* substantial numbers of drivers are making voice calls; (Available
evidence, part of main response to Option A);
Level 3: Strong quantitative evidence:
Despite varied studies of drivers’ voice-phoning, there is still
uncertainty as to the extent of impact on safety. Strong quantitative
evidence is lacking on the extent of driver texting and other more
recent forms of electronic distraction.
NHTSA responses to electronic distractions including action or research
to counter this trend:
Option A: No self-initiated response:
* NHTSA’s main response is a decision not to self-initiate action or
research on countermeasures, at least not at this time. (Selected
response)
Option B: “Starting point” actions:
* NHTSA is not initiating forums, policy proposals, or a vision for
this issue at this time[B].
Option C: Research to explore or evaluate approaches to action or to
stimulate action; early interventions:
* Based on a SAFETEA-LU requirement, NHTSA is funding two demonstration
programs to counter teen driver distraction; one of these focuses
specifically on electronic distractions; (Selected response)
Option D: Mid-level interventions:
* Based on a statement in a congressional conference report, NHTSA (1)
contracted for and published a literature review, and (2) provided it
to DOT regions for distribution to the states;
* The review concludes that driver phoning may increase risks by a
factor of 4 and points to GDL cell bans as a possibly effective
countermeasure[C]. (Selected response)
Option E: Stronger interventions:
* NHTSA has no regulatory authority for portable devices like cell
phones;
* Without congressional authority, NHTSA cannot (1) make new grants to
states for distraction-safety activities or (2) make other grants to
states contingent on states passing laws related to distractions.
Note: The arrow indicates NHTSA's current main response to the driver
distraction trend, which is not to self-initiate research or action to
counter it at this time. Other highlighted boxes indicate other current
or recent responses, based on congressional directives.
[A] Besides providing evidence on safety impact, research has examined
the process of distracted driving and ways to study it; for example, a
2008 report NHTSA contracted for discussed two earlier studies' results
on driver willingness to engage in distracted behaviors (Lerner,
Singer, and Huey 2008, citing Lerner and Boyd 2004 and Lerner and
Balliro 2003).
[B] A table in Lerner, Singer, and Huey (2008) outlines "possible
countermeasures to address concerns" on distracted driving. These
concerns are based on the earlier studies cited in note a, above. The
outline could provide background for planning or conceptualizing new
starting point actions (Option B) or new research on countermeasures
(Option C). NHTSA's Internet forum and pilot teen-driver education
campaign (Smart Drivers Just Drive) on distractions were both
discontinued.
[C] GDL refers to graduated driver licensing programs that involve
stages; some programs ban cell phone use for novice drivers or 16-and
17-year olds.
[End of figure]
Additionally, although quantitative data on driver texting and the use
of other new electronic distractions are more limited than data on
driver voice-phoning, a number of "early signals" indicate that such
behaviors also increase safety risks:
* Two driving-simulator studies--one of driver text messaging, the
other of drivers' using MP3 players--indicate that these distractions
have negative safety effects (Chisholm, Caird, and Lockhart forthcoming
and Hosking and others 2006).[Footnote 44]
* A lead investigator on the 100-car study and a similar ongoing study
of teen drivers recently stated that texting, iPod and MP3
manipulation, and Internet interaction are (like cell phone dialing)
riskier than talking on a cell phone (Dingus 2007).
* In a 2008 self-report survey of cell phone users, a slight majority
of 20-to 29-year-old respondents said they text message while driving,
as did more than one-third of respondents in their thirties and over a
fourth of those in their forties (Vlingo 2008).[Footnote 45]
These early signals are Level 1 evidence in figure 13.
Looking to the future, other early signals suggest that more
distracting devices and greater amounts of driver use may further
increase risks:
* From now through 2015, more complex devices may emerge. Industry
representatives told us they "do not see technology slowing down" in
the coming years. They expect continued evolution of electronic devices
in a similar direction: increasing complexity of devices, more
applications, detailed screens, and fast proliferation.
* From 2015 through 2020, "cohort effects" may significantly increase
the percentage of drivers using devices. This would occur if young
drivers continue texting and middle-aged drivers continue voice calling
as they age and if new cohorts of teen drivers text or use newer
complex devices at levels similar to or higher than today's
teens.[Footnote 46] (This assumes that technology or behavior changes
do not alter these anticipated outcomes.)
Now and in the future, teens may have the highest risks. The recent
NHTSA-sponsored literature review, which noted that cell phone use is
increasing, cited findings that younger, and in some cases novice,
drivers are "leading the way" in using various new devices and that the
combination of distraction and lack of "fully developed driving skills"
suggests accelerating risks for this group (Ranney 2008, 15).
Additionally, driver use of portable phones with touch screens can now
be facilitated with dashboard holders with swivel mounts for landscape
and portrait viewing. Motorcycle helmet equipment is also now available
to facilitate phoning while riding (fig. 14).[Footnote 47] Finally,
wireless Internet is becoming available in cars that will become "a
moving WiFi hotspot with Internet access" (Newman 2008).
Figure 14: Dashboard Mounts and Helmet Equipment for Phoning:
[Refer to PDF for image]
This figure contains illustrations of dashboard mounts and helmet
equipment for phoning.
Source: ProClip (portable device holder); Cardo Systems, Inc. (helmet).
Note: Dashboard mounts (left) accommodate a smart phone or PDA
(personal digital assistant), cell phone with screen, GPS device, and
MP3 player.
[End of figure]
NHTSA has acted on two directives shown in figure 13 Options C and D:
* Option C: With SAFETEA-LU, the Congress adopted a requirement that
DOT fund at least two demonstration programs for mitigating the effect
of distracted, inattentive, and fatigued driving.[Footnote 48] NHTSA
told us that as of July 2008, it was making final awards for
subcontracts for two teen-driver distraction projects, one focusing
specifically on electronic distractions.
* Option D: In response to a statement in the conference report
accompanying DOT's fiscal year 2006 appropriation--which called for "an
effort to consolidate current knowledge on driver distraction for use
by policymakers …[with the purpose of assisting] state and local
governments to formulate effective policies, regulations and laws"--DOT
contracted for and published a literature review on driver
distractions. DOT told us that it also provided for its distribution to
states.[Footnote 49]
These activities involve NHTSA's acting as a partner with the states.
However, at this time, NHTSA's main response to the electronic driver
distraction issue is a decision not to self-initiate either research
specifically aimed at countering such distractions or other actions--
Option A in figure 13 (although NHTSA is studying the safety impact of
distracting behaviors, including driver use of electronic devices).
NHTSA has not yet implemented other suggestions or directives that
government stakeholders, at the federal and state levels, have made.
Two of these correspond to governance Option D in figure 9:
* The Congress directed DOT to develop uniform guidelines for state
programs, including those aimed at reducing accidents resulting from
unsafe driving behavior, such as the use of distracting electronic
devices.[Footnote 50] NHTSA told us that it will not begin drafting
such guidelines until the SAFETEA-LU demonstration projects are
completed.
* The Governors Highway Safety Association (GHSA) asked the federal
government to fund a comprehensive media campaign to educate the public
about the dangers of distracted driving (including but not limited to
the dangers of electronic distractions) and how to manage distractions
(GHSA 2003).[Footnote 51] NHTSA told us that it is not planning to
conduct such a campaign at this time.
An additional suggestion for possible research is contained in the
congressional conference report that led NHTSA to conduct the
literature review. The conference report suggested that the results of
the literature review could help "focus the federal research effort in
the most productive directions." The review stated that a useful
research direction would be to evaluate the impact of state laws--and
suggested that a promising approach to countering distractions might be
state graduated drivers licensing (GDL) programs that ban cell phone
use by new drivers (Ranney 2008, 18).[Footnote 52] NHTSA has not begun
related research. (We discuss the suggested focus on a GDL cell ban in
terms of possible research on state laws in appendix IX.)[Footnote 53]
NHTSA officials did mention that crash avoidance technologies might
mitigate distractions, but evolving crash avoidance and electronic
distraction technologies--or other trends--might interact in other
ways. (Examples of unintended consequences from trend interactions are
in appendix VIII.)
In contrast to NHTSA's position, various industry groups are
supporting, suggesting, or initiating responses to the driver-
distraction trend. CTIA--The Wireless Association® (representing cell
carriers) supports state bans on (1) text messaging for all drivers and
(2) all cell phone use by provisional or novice drivers, except in
emergencies. Telecommunications Industry Association (TIA) officials
told us that the development of new technologies might mitigate this
problem and that it is working to develop new technical standards to
help speed the progress of such technologies. The Consumer Electronics
Association (CEA) has developed driver education materials.[Footnote
54]
Several safety advocates told us they are concerned that DOT is not
aggressively addressing this trend. Finally, while policy on highway
safety should not necessarily be driven by public opinion polls, recent
polls indicate that a majority of those questioned believe that driver
use of electronic devices is significantly more distracting than other
activities and favor a ban on driver texting.[Footnote 55] A poll of
teens reported their opinion that driver texting and other handheld
device use (other than talking on a cell phone) is second only to
alcohol in its negative impact on road safety (Ginsburg and others
2008).
Decision Making on High-Clockspeed Trends: NHTSA Criteria and Other
Approaches:
NHTSA officials state that they have specific criteria and that they
have applied them consistently to decisions on the two trends we
examine in this report, as well as to other decisions about highway
safety. According to NHTSA officials,
* Their top criterion is evidence that the safety problem is large (of
a size akin to lack of seatbelt use or alcohol use or lack of helmet
use), and they have more evidence of this with respect to the "old"
problems addressed by crash avoidance, such as cars running off the
road or rear-ending other cars, than the new problem of electronic
driver distractions.
* Their second main criterion is having a promising countermeasure to
address the safety problem, and they have this for in-vehicle crash
avoidance technologies that address issues, such as running off the
road, but not in the case of electronic driver distractions--although
NHTSA officials say they would work on exploring new strategies to
address a problem once it is clear that the problem is large.
* In addition, NHTSA officials say that they take many other factors or
criteria into account, as appropriate (costs, impacts on mobility,
likely success in addressing the issue, diminishing returns from
working on old problems, and so forth).
When we questioned NHTSA officials further about their current choice
of the "no self-initiated response" option for electronic driver
distractions, they listed decision criteria that represented possible
constraints and that did not include considering the trajectory of this
trend or its possible future phases.[Footnote 56] They also said their
decisions were made without using a formal risk management approach.
[Footnote 57]
The approach to decision making that NHTSA described to us can result
in its overlooking the potential significance of some high-clockspeed
trends and possible ways of shaping or countering them. For example,
the first main criterion--evidence of a large safety problem--does not
allow for "worsening" problems (from high-clockspeed change) or
difficulties in obtaining quantitative data on new problems
(uncertainty). The second main criterion does not allow for exploratory
research aimed at finding a promising countermeasure (and reducing
uncertainty) until a problem has reached major proportions and been
documented. Anticipatory risk management that embraces the uncertain
environment that NHTSA faces is lacking in NHTSA officials' stated
decision-making framework.
Furthermore, NHTSA may be inconsistent in selecting additional criteria
to apply to new trends. For example, in the case of crash avoidance,
NHTSA compared the safety gains that might be made using an old versus
a new solution (that is, there was a sense of diminishing returns for
crashworthiness, which might be reversed with a new technology) but
this criterion may not be applied to all trends. Finally, although
NHTSA officials explained their criteria to us, they had not documented
them.
Foresight literature and our discussions with experts suggest a number
of approaches to (1) reducing uncertainty to support decisions and (2)
making decisions when substantial uncertainty exists. Both might be
relevant to decisions on high-clockspeed trends. These varied
approaches include the following:
* Expanding existing networks as new trends develop. Some industries
may have been of limited relevance to old problems or long-established
solutions but might provide early signals about new trends, such as
information about new technologies "in the pipeline." Such information
may help an organization anticipate new phases of a trend and
potentially changing impacts (such as the changing safety impacts of
driver distractions).
* Anticipatory investing in research on ways to shape or counter new
trends. Prioritizing research responses to a new trend can produce
information that reduces uncertainty about possible future action,
should events prove to justify this. Anticipatory or preparatory
research can facilitate timely action at a later date.
* Clarifying criteria for action responses intended to shape "old"
versus potential new solutions or to counter "old" versus new problems.
A number of different criteria may be applicable in different
situations or may be appropriately balanced or combined in any one
situation. Possible criteria include (1) the relative sizes of old and
new problems; (2) diminishing returns from investments in old problems
or old solutions, compared to potential returns from new investments;
and (3) the cost-benefit ratio anticipated for a new research or action
option.
* Using a formal risk management approach that involves an anticipatory
perspective. Risk management frameworks suggest outlining various
alternative responses to new trends and specifying the risks associated
with pursuing them. For example, the risks of delaying action are
higher--and an anticipatory approach may be more appropriate--when one
or more of the following apply: (1) the trend in question has high
clockspeed, (2) prospective interventions require long lead times, (3)
high-stakes outcomes would be likely to be affected, and (4) a delayed
intervention might allow negative developments to gain a foothold,
making impacts difficult to contain or reverse.[Footnote 58] In some
cases, delaying research may rule out subsequent timely action.
[Footnote 59]
* Considering adaptive strategies that avoid risky interventions based
on a single, assumed future course of events. Adaptive strategies are
designed to change or evolve in response to new information as it
becomes available. For example, an agency might plan back-up responses
("plan B" or "plan C") for use in the event that certain key ("plan A")
assumptions fail. Another alternative that might be feasible in some
cases is to choose an option that is robust in terms of its ability to
work reasonably well across many--perhaps all or almost all--
alternative future developments identified (see Popper, Lempert, and
Bankes 2005 and Dewar 2006).
Developing new quantitative Level 3 evidence on the impact of new
trends on safety could, if successful, reduce uncertainty, as discussed
in the following section.
DOT Faces the Challenge of Devising Timely Measures of Change over
Time:
Developing strong quantitative evidence on the safety impact of a new
trend--especially 21st century technology-based trends such as post-ESC
crash avoidance technologies or electronic driver distractions--can be
challenging. Long-standing methods may not be suited to studying new,
high-clockspeed trends. However, successfully developing improved
evidence can help decision makers by reducing a key source of
uncertainty (see figure 15).
Figure 15: Exercising Foresight by Addressing High-Clockspeed Trends,
with Evidence Development Highlighted:
[Refer to PDF for image]
This figure depicts the following information, with the second
highlighted:
Deciding and responding:
Deciding how to respond to or shape each potentially significant high-
clockspeed trend, considering available evidence and uncertainties.
[Highlighted: Developing evidence:
Providing additional evidence on the effects of high-clockspeed trends,
to reduce uncertainty (may include devising new data systems or
analysis methods)].
Communicating:
Communicating effectively with the Congress and others about high-
clockspeed and other trends, agency responses, and policy implications.
Source: GAO.
[End of figure]
DOT has used both long-standing and newer methods to develop evidence
on the safety impacts of crash avoidance technologies and electronic
driver distractions. None are suited to producing strong quantitative
evidence on safety impacts of high-clockspeed trends.[Footnote 60] As
defined here, such evidence combines three key characteristics:
technical adequacy, timeliness, and the measurement of change over
time.
* Technical adequacy includes validity, reliability, and
generalizability, which are established methodological strengths for
data and evidence in general and also relevant to producing evidence
for foresight.[Footnote 61]
* Timeliness is generally recognized as a strength of policy-relevant
data.[Footnote 62] Evidence that lags behind trends can result in
decisions that are of fading relevance or that apply only to
technologies or behaviors now replaced by newer ones. However,
timeliness is a relative concept. In slow-changing areas, still-
relevant data may have been developed in a 3-year project that was
completed 5 or more years ago. But for high-clockspeed trends, timely
evidence means recent evidence--for example, data or tests conducted
during the previous year. More time-consuming methods of data
collection or assessment would not be adequate.
* Measurement of change over time means tracking the trajectory of
trends or their changing impacts and is related to timeliness. The
measurement of change at relatively frequent intervals is an important
characteristic of evidence on the impacts of high-clockspeed trends. If
not updated at frequent intervals, the evidence will not be timely.
Information on trajectories and currently changing impacts can suggest
directions of future developments and can help shape forward-looking
decisions and policies. However, tracking high-clockspeed trends
requires repeated data collections at relatively frequent intervals.
Unless methods are economical, repeated applications can be costly.
DOT has used both long-standing and newer methods in its attempts to
develop evidence on post-ESC crash avoidance technologies and
electronic driver distractions, but none combine all three
characteristics discussed above.[Footnote 63]
Existing Methods for Assessing Crash Avoidance Do Not Combine Technical
Adequacy with Timeliness and Measurement of Change:
Using NHTSA's NCSA crash datasets to test the safety effectiveness of
new crash avoidance technologies can produce technically adequate data,
as indicated in table 1. Notably, NHTSA earlier analyzed multiyear
crash datasets to compare new or late-model cars with and without ESC
and demonstrated a substantial ESC safety benefit.[Footnote 64]
However, NHTSA officials told us that this approach does not produce
timely assessments of new crash avoidance technologies. Specifically,
NHTSA estimated that, if it were to use its crash datasets, 5 to10
years would be needed to assess each technology. This would delay
action to improve safety; moreover, by the time such assessments were
issued, the tested technologies might have been replaced by newer
versions. Lengthy assessments also make it difficult to measure change
over time, because they essentially rule out repeated updates that
track ongoing improvements in safety effectiveness.
Table 1: Methods for Assessing the Safety Impacts of Evolving Post-ESC
Crash Avoidance Technologies Rated by Three Key Characteristics:
Assessment method: Long-standing: Analysis of crash data collected over
multiple years (NCSA datasets);
Characteristic applied in assessment: 1. Is technically adequate: Yes;
Characteristic applied in assessment: 2. Is timely: No[A];
Characteristic applied in assessment: 3. Measures change over time:
No[A];
Assessment combines all three characteristics: No.
Assessment method: Newer: Field operational tests with complex
technologies (measures driver performance, not crashes)[B];
Characteristic applied in assessment: 1. Is technically adequate: Not
fully[C];
Characteristic applied in assessment: 2. Is timely: Yes;
Characteristic applied in assessment: 3. Measures change over time: No;
Assessment combines all three characteristics: No.
Source: GAO.
[A] It would take several years to amass sufficient data to test each
new technology that evolves, thus ruling out the possibility of
tracking changes as they occur from year to year.
[B] Field operational tests involve NHTSA or another DOT agency or
contractor outfitting a limited number of cars with instruments such as
cameras, sensors, and recorders to track driving experiences.
[C] There is some disagreement on whether studies involving cameras,
sensors, or other technologies for intensively recording data change
driver behavior. Some researchers told us that when drivers know they
are being closely observed, they may not behave as they otherwise
would; however, advocates of naturalistic studies said that many
drivers become used to cameras and disregard them.
[End of table]
Using NHTSA's crash datasets to assess new in-vehicle technologies
would involve lengthy time periods because a single year of new-car
crash data from these datasets is not sufficient for making meaningful
comparisons of crashes with and without the technology:
* initially, only some new cars are equipped with a particular new
technology and newly purchased cars represent a relatively small
percentage of cars being driven, and:
* rather than encompassing all serious crashes, NHTSA's ongoing crash
datasets are limited to either (1) the fraction of crashes that are
fatal (the FARS dataset) or (2) a sample of more broadly defined
serious crashes (the National Accident Sampling System's
Crashworthiness Data System or NASS CDS).[Footnote 65]
Additionally, states may take several months to collect, process, and
report data to NHTSA, after which NHTSA researchers further process and
analyze them. As a result, NASS CDS data are typically available in
draft form, 9 months after the quarter in which a crash occurred.
To achieve more timely assessments of crash avoidance technologies,
NHTSA used FOTs. Although FOTs produce some useful data, their small
sample size limits results and means that driver performance measures,
not crashes, must be used as the outcome. Notably, for LDW and FCW, the
FOTs were not extensive enough for analysts to specify the magnitude or
level of safety benefit that each technology provides. Recently, NHTSA
combined FOT data with other information on the frequency of various
types of crashes to project quantitative levels of "safety
effectiveness." (In the previous section, we classified these
projections as Level 2 evidence; the projected benefits were much
smaller than demonstrated for ESC.) Additionally, a NHTSA official told
us that FOTs using complex tracking technologies are relatively
expensive, so from a practical cost perspective, it would be difficult
to repeat them each time a new technology (or an improved version of
recent technologies) is introduced.
Existing Methods for Assessing Distractions Do Not Combine Technical
Adequacy with Timeliness and Measurement of Change:
DOT encountered technical adequacy problems with both long-standing
methods used to measure new driver distractions (see table 2):
* Crash datasets maintained by NCSA miss an unknown number of precrash
distractions, since these may be hidden from police and investigators
or may not be recorded on police accident reports.[Footnote 66]
* Roadside observations of driver phoning, as part of the National
Occupant Protection Use Survey (NOPUS) that NHTSA developed earlier to
measure seatbelt use, are likely to miss instances of hands-free
phoning when drivers use earpieces or speakerphones.[Footnote 67]
Table 2: Methods for Assessing the Safety Impacts of Evolving
Electronic Driver Distractions Rated by Three Key Characteristics:
Assessment method: Long-standing: Analysis of data in NCSA crash
datasets;
Characteristic applied in assessment: 1. Is technically adequate: No;
Characteristic applied in assessment: 2. Is timely: Potentially,
yes[A];
Characteristic applied in assessment: 3. Measures change over time:
Potentially, yes[A];
Assessment combines all three characteristics: No.
Assessment method: Long-standing: Surveys of driver behaviors:
observational and landline;
Characteristic applied in assessment: 1. Is technically adequate: No;
Characteristic applied in assessment: 2. Is timely: Yes;
Characteristic applied in assessment: 3. Measures change over time:
No[B];
Assessment combines all three characteristics: No.
Assessment method: Newer: Naturalistic studies using cameras, sensors,
and other tracking technologies[C];
Characteristic applied in assessment: 1. Is technically adequate: Not
fully[D];
Characteristic applied in assessment: 2. Is timely: Potentially, yes;
Characteristic applied in assessment: 3. Measures change over time: No;
Assessment combines all three characteristics: No.
Source: GAO.
[A] Because the population of drivers using portable electronic devices
is extensive (and includes drivers of new and older cars), analyses
focused on electronic driver distractions could be conducted using
fewer years of data than would be required to assess new-car safety
technologies.
[B] To the extent that hands-free phoning and cell-phone-only
households are increasing, estimated trends would be invalid if based
on either roadside observations that missed hands-free phoning or
landline phone surveys that missed cell-phone-only households.
[C] Depending on the size of the study, near misses, not crashes, may
be used as the outcome.
[D] Researchers disagree on whether subjects in such studies may change
their behavior because they know they are being observed.
[End of table]
Newer methods of developing evidence on this trend are also not fully
adequate. NHTSA told us that it is attempting to identify new
technology that might improve future roadside observations--that is,
technology that might be able to detect (from the roadside) whether a
cell phone transmission is in progress. Most recently, NHTSA adjusted
its observations-based estimate, using responses in a telephone survey,
based on a sample of landline households. Key responses concerned how
often drivers used handheld versus hands-free phones. One issue is that
this landline telephone survey omits cell-phone-only households.
According to a survey contractor we consulted, the omitted cell-phone-
only households represented about 15 to 18 percent of the adult
population and 20 to 30 percent of groups such as young adults--and
omitting such households can yield biased estimates of behavior such as
technology adoption.[Footnote 68]
To better assess the safety impact of driver distractions (as well as
to understand other aspects of how crashes occur), DOT pioneered
naturalistic studies in which video cameras, other sensors, and
recorders are installed in participating drivers' personal cars. Each
car is then tracked over time, so that recordings provide analysts with
volumes of detailed video and other data. Although smaller-scale
studies of this type, such as the 100-car study and an ongoing study of
teen drivers, can provide important insights into safety processes,
results for crash risk are limited because of the small size of these
studies.[Footnote 69]
Larger-scale studies of this type can produce more technically adequate
estimates of crash risk. TRB, in cooperation with DOT, is designing a
2,500-car study to measure many aspects of safety and driver behavior,
including driver distraction. DOT officials anticipate that assuming
the 2,500-car study is successfully fielded, it is likely to:
* provide a useful snapshot of the nature of distracted driving
behavior and the magnitude of the effect that distracted driving has on
actual crashes as of 2009 to 2010[Footnote 70] and:
* test the validity of substituting near misses for actual crashes in
analyses of specific kinds of driving or crash situations).[Footnote
71]
Fieldwork on the 2,500-car study will begin in 2009 or 2010.[Footnote
72] However, equipping large numbers of vehicles with cameras, sensors,
and other equipment is costly and, therefore, in the opinion of a TRB
official and a NHTSA official, it will be a challenge to fund similar
studies in the near future.[Footnote 73]
Given experience with naturalistic studies designed to date, the high
cost of larger-scale versions of naturalistic studies appears to be
limiting. The future of smaller-scale naturalistic studies may depend,
in part, on how well, or under what conditions, near misses are shown
to be valid surrogates for actual crashes. Additionally, estimates of
the impact of driver distractions based on naturalistic studies may not
be directly comparable to estimates of the impact of elevated driver
blood alcohol content, based on crash datasets, because naturalistic
studies exclude some dangerous drivers who are included in crash
datasets.
The newer methods tried to date have the potential to meet the
timeliness criterion but have not so far been suited to measuring
change over time. Specifically, it seems unlikely that large-scale
naturalistic studies such as the one described above will be repeated
and, therefore, it is also unlikely that studies like this could be
used to track changing impacts of evolving electronic distractions over
time (that is, meet the criterion of tracking change).
Meeting the Evidence-Development Challenge for High-Clockspeed
Technology Trends:
Developing high-quality evidence on the impacts of high-clockspeed
technology-based trends is considerably more difficult than continuing
to study long-established or slow-changing safety technologies or
behavioral issues. This is because the criteria above--in particular
timeliness and tracking change over time--are more important for high-
clockspeed trends. In addition, new kinds of safety technologies may be
sufficiently different from older technologies that previously
developed assessment methods cannot be applied.[Footnote 74] In
addition, methods must be flexible enough to work across multiple
stages of an evolving trend. Technologies or behaviors may change as
trends evolve, so that--in at least some cases--methods that worked at
an early generation or stage may not be suited to studying a later one.
[Footnote 75]
Some examples of the kinds of new data systems or analyses that might
address these issues were suggested by experts we talked with as well
as the literature that we reviewed, including a recent NHTSA-sponsored
report. We have not determined the costs of such systems. The systems
and analyses include:
* A new research tracking system, with equipment built into new cars--
provided that appropriate privacy protections are designed and
incorporated. OnStar™ tracks crashes for cars equipped with OnStar™.
Recent General Motors models with OnStar™ can track all crashes
involving a certain level of impact (with and without airbag
deployment). Crash data are sent wirelessly to a remote location, where
they are analyzed by OnStar™ researchers to determine if ways may exist
to design safer vehicles. Logically, similar technology might be used
to track crashes for most or all new cars in the future.[Footnote 76]
* Use of new or developing technologies to track driver use of
electronic equipment--in general and in relation to crashes--provided
that, where needed, appropriate privacy protections are designed and
incorporated. Varied technologies now available or under development
can potentially (1) help roadside observers detect cell phone use in
passing cars or (2) otherwise allow researchers to track driver use of
cell phones, including whether they were phoning just before a crash
(Brennan, Adi, and Campbell 2008). NHTSA officials told us that NHTSA
is investigating the former. The latter is more speculative in that it
would require installing special equipment in each car to be tracked;
however, it has the advantage of being able to distinguish driver use
of a device from passenger use and voice transmissions from data
transmissions.
* Statistical models that combine varied data and can support simulated
premarket testing. Recent conference papers and a DOT-issued report by
a NHTSA contractor outline the possibility of models that would detail
precrash and crash descriptors (starting with precipitating events such
as a car's drifting out of its lane, proceeding to the driver's
attempting to avoid a crash, and then to the first harmful event, and
so forth). Such models would allow researchers to combine data from
NHTSA's crash datasets and naturalistic data to, for example, define
scenarios for use in designing test-track assessments of new crash
avoidance technologies (Burgett, Srinivasan, and Rangunathan 2008).
DOT Faces the Challenge of How and When to Communicate Information on
Trends to the Congress:
A final challenge for DOT is to effectively inform stakeholders,
including the Congress, about the implications of high-clockspeed
technology trends (see figure 16). These trends could have an impact on
key decisions made during the reauthorization of funding for surface
transportation programs. While DOT recently developed a framework to
prompt deliberation by the Congress and other stakeholders, this
framework and DOT-wide planning and accountability materials are not
designed or intended to provide a long-term view or comprehensive
analysis of trends that could affect highway safety in the future--
including evolving crash avoidance technologies and rapidly changing
electronic driver distractions--and their interactions and implications
for the years ahead.
Figure 16: Exercising Foresight by Addressing High-Clockspeed Trends,
with Agency Communication Highlighted:
[Refer to PDF for image]
This figure depicts the following information, with the third
highlighted:
Deciding and responding:
Deciding how to respond to or shape each potentially significant high-
clockspeed trend, considering available evidence and uncertainties.
Developing evidence:
Providing additional evidence on the effects of high-clockspeed trends,
to reduce uncertainty (may include devising new data systems or
analysis methods).
[Highlighted: Communicating:
Communicating effectively with the Congress and others about high-
clockspeed and other trends, agency responses, and policy
implications].
Source: GAO.
[End of figure]
In addition, DOT has not synthesized the results of its various
foresight efforts in order to show how overall trends might impact
highway safety in 2020 and beyond. Some of DOT's own practices and
other models from the United States and abroad might provide strategies
for communication specifically intended to inform and update the
Congress about trends with the purpose of enhancing or supporting
foresight in decision making.
DOT Provides Limited Information on Highway Safety Trends in
Reauthorization Framework and GPRA Materials:
We have reported that an agency's assessment of trends or factors that
are external to its environment may help the Congress in judging the
likelihood of achieving strategic goals and actions needed to meet
those goals (GAO 1997, 18). Information on external trends or factors
can also enable an agency to explain why performance goals are not
met.[Footnote 77] And according to one expert, a detailed analysis of
long-range trends can provide support for government, including
congressional, decision making. Available mechanisms for the provision
of such information include the reauthorization of funding for programs
and GPRA-related documents, such as the strategic plan and performance
and accountability report.
According to DOT officials, DOT formally communicated to the Congress
in advance of reauthorization, on July 29, 2008, putting forth a
framework intended to spur local, state, and federal debate about the
way U.S. transportation decisions and investments are made (DOT 2008b).
According to DOT, "ongoing demographic, economic, technological,
political and institutional trends have major implications for our
Nation's transportation system now and well into the future." With
respect to safety, DOT "heavily emphasize(s) the potential of various
crash prevention technologies to significantly reduce highway
fatalities." The document does not provide information on trends beyond
the references above that could be used by the Congress.
In terms of a future horizon, DOT's strategic plan acknowledges future
long-term trends with the inclusion of VMT and demographic projections
for periods ranging from 2030 to 2050 but does not address how other
new and emerging trends might affect highway safety beyond the plan's 6-
year projection. DOT has met the GPRA requirement that agency plans
cover a period of not less than 5 years forward from the reporting
fiscal year.
An Overall Context That Considers Multiple Trends Is Lacking:
According to DOT, "interrelated facets of the transportation problem...
must all be considered together in evaluating the complete costs and
benefits of transportation measures."[Footnote 78] We found limited
attention in departmentwide materials to laying out such interactions
and their implications for the future, either in recent, focused
reviews conducted in response to congressional interest or in planning
and accountability materials issued on a continuing basis.[Footnote 79]
The recent DOT 1.0 Fatality Rate Goal report cited the example of aging
of baby boomers and rising fuel prices as they might affect motorcycle
fatalities in the near-term future (DOT 2008a).[Footnote 80] It did not
explicitly discuss how trends might interact to either positively or
negatively affect highway safety in a future beyond 2011.
The DOT 1.0 Fatality Rate Goal report also did not include information
on the trend of electronic driver distractions, which is important for
understanding the interaction of multiple trends. Relevant information
is included in a NHTSA-sponsored review of literature on driver
distractions initiated in response to the conference report for the
2006 DOT appropriations bill. The NHTSA-sponsored review discussed cell
phones as "the contemporary icon of driver distraction," summarized the
results of existing studies, and acknowledged that problems may
increase from (1) the continually increasing number of cell phone users
and (2) the "secondary use" of cell phones for activities other than
talking, such as text messaging by teenage drivers who may lack fully
developed driving skills (Ranney 2008). As a result, "we may expect to
observe a synergistic acceleration in the resulting safety problem."
Despite being in the NHTSA-sponsored review, such observations were
excluded from the DOT 1.0 Fatality Rate Goal report.
DOT's GPRA-related documents focus their discussion on well-established
trends, such as the aging of the U.S. population, and do not address
how interrelated trends related to the transportation system as a whole
could affect highway safety. In addition, they are not designed to
discuss the implications of trends (whether well-established or in the
early stages of development) in a detailed fashion. GPRA and related
OMB guidance do not require such specificity in these materials.
In the past, DOT initiated major efforts aimed at understanding
multiple trends as part of the broad scope of change affecting
transportation and considered highway safety within this context. For
example, the Secretary of DOT initiated a series of "2025 Visioning
Sessions" with hundreds of transportation stakeholders around the
country on how transportation would be likely to change in the next 25
years. The results of such sessions provided input for two reports DOT
issued in 2000: The Changing Face of Transportation and Transportation
Decision Making: Policy Architecture for the 21st Century (DOT 2000a,
b). The first report outlined major trends expected to unfold in the
next 25 years, and the second provided a framework to support decision
making in transportation.[Footnote 81]
More recently, the Office of the Secretary convened policy salons
intended to educate DOT officials and staff on broad transportation
trends, with some attention to highway safety developments, such as
GM's technology plans for the future, but this effort no longer exists
and no report was issued analyzing the results of these sessions for
use by Congress and stakeholders outside the Department.[Footnote 82]
Future-Oriented Efforts at DOT Are Not Synthesized for Use by the
Congress:
We found that while the DOT 1.0 Fatality Rate Goal report provided a
valuable description of current initiatives and strategies, information
from future-oriented or foresight efforts throughout DOT has not been
fully captured by this report. Such efforts have produced information
to meet planning and other mission-specific needs of various
administrations. DOT has not synthesized information from these various
documents or presented such a synthesis to the Congress. Table 3 gives
examples of DOT future-oriented efforts.
Table 3: Examples of Foresight across DOT:
Type of foresight activity: Vision development;
DOT administration and effort: RITA, A Vision for Transportation[A];
Horizon: 2030;
Purpose or target: Context-setting for DOT transportation research,
development, and technology investments;
Highway safety addressed?: Enunciates principles; highway safety and
relevant technology developments described in context of overall
transportation system.
Type of foresight activity: Trend analysis, scenario development, and
operational implications;
DOT administration and effort: FMCSA, Motor Carrier Study 2025[B];
Horizon: 2025;
Purpose or target: FMCSA planning;
Highway safety addressed?: Addresses highway safety in terms of FMCSA
mission and provides detailed analysis of technology and other trends
that could affect safety in the motor carrier industry.
Type of foresight activity: Use of experts to identify future trends in
transportation;
DOT administration and effort: FHWA, Advanced Research "Think Tank"
Forums[C];
Horizon: 2050 as context;
Purpose or target: Developing an advanced research plan;
Highway safety addressed?: Not explicitly; FHWA officials noted that
the overall effort prompted subsequent discussion at FHWA on the need
for and development of a conceptual framework for highway safety at
DOT.
Type of foresight activity: Technology scanning[D];
DOT administration and effort: FHWA, International Highway Technology
Scanning Program;
Horizon: Ongoing;
Purpose or target: Disseminate best practices;
Highway safety addressed?: Addresses various facets of highway safety.
Type of foresight activity: Unintended consequences;
DOT administration and effort: NHTSA, Human Factors Forum on Advanced
Vehicle Safety Technologies[E];
Horizon: January 2007 workshop;
Purpose or target: Identify research priorities through interaction
with stakeholders;
Highway safety addressed?: Elicited stakeholders' comments on
unintended consequences of crash avoidance technologies and prioritized
items for future research.
Source: GAO.
Note: DOT told us that another example of a foresight activity is a
current BTS effort to revisit its earlier long-term forecasts, review
their accuracy, and update them.
[A] DOT 2008c.
[B] Internal report prepared for FMCSA.
[C] These forums are described in Asmeron and McRae 2006.
[D] As defined by FHWA, scanning features the formation of expert teams
(managers and specialists in a particular discipline) that travel
abroad to consult with foreign counterparts in other countries where
advances in transportation relevant to the United States are being
made. Scan team members typically represent FHWA, state departments of
transportation, local governments, transportation trade and research
groups, the private sector, and academia. When a scan is completed,
team members evaluate findings and develop a comprehensive report that
is circulated throughout the U.S. highway transportation community. The
team also develops an implementation plan that summarizes its strategy
for implementing the most significant and promising technologies and
policies the scan identified. According to FHWA, to accelerate early
implementation activities, the scan program supports teams with
implementation expertise and funding when they return to the United
States.
[E] These issues are discussed in Wochinger and others 2008.
[End of table]
One example, in 2006, was FMCSA's commissioning a study to consider
likely forecasts of future industry scenarios in 2025 in order to learn
how its current programs, policies, management, and operations could be
adjusted to better achieve safety objectives in the next two decades.
[Footnote 83] With historical trends and forecasts providing the basis
for scenario development, the resulting study reviewed current FMCSA
programs and initiatives to determine whether the administration was
well positioned to respond to some of the scenarios. For example, the
study assessed the extent to which CSA 2010, a new operational model
for FMCSA intended to reduce crashes, fatalities, and injuries related
to commercial motor vehicles will affect various scenarios ranging from
driver fitness to technology improvement. And after describing a
technology-related scenario, the study discussed implications for FMCSA
operations, such as human capital and future operational needs and
assigned responsibility to FMCSA's Office of Administration in this
regard.[Footnote 84]
Activities such as these are carried out on a one-time basis for a
specific purpose and are generally not formally transmitted to the
Congress. We did identify some examples in which DOT carried on the
development of foresight information in a continuous fashion, but as is
the case with the activities we have described, they have not been
synthesized for use by the Congress. In addition to FHWA's ongoing
development of scans targeting specific issues, NHTSA has developed an
ongoing process for eliciting comment on the inclusion of future
technologies in NCAP. FHWA's International Technology Scanning Program
is characterized by an assessment of innovations and practices abroad;
it began in the early 1990s. The program has published the results of
scans of innovative technologies and practices in other countries that
could significantly benefit U.S. highway transportation systems. Such
scans could lead to the development of new strategies. For example,
after a scan of highway safety information systems in other countries,
FHWA safety officials developed a white paper detailing a strategy and
steps for further action, including the involvement of multiple
stakeholders (FHWA 2006c).
Another scan is intended to survey policy approaches that could improve
U.S. roadway safety for older users and, consequently, all road users.
[Footnote 85]
Like FHWA, NHTSA has attempted to stay abreast of international
developments, such as those with implications for establishing fatality
reduction goals and managing trends related to crash avoidance. For
example, in 2007, the director of traffic safety in Sweden and chair of
the European New Car Assessment Programme (Euro NCAP) briefed DOT
officials on the Swedish approach to fatality reduction known as Vision
Zero, as well as trends in Euro NCAP. More generally, officials told us
they attend conferences and network with key stakeholders to gain an
understanding of technology and behavioral developments relevant to
highway safety.[Footnote 86] NHTSA officials have disseminated results
in conference presentations and articles that suggest options for
action. In addition, NHTSA has communicated with stakeholders on how it
plans to address the trend of in-vehicle crash avoidance. For example,
NHTSA developed a January 2007 proposal published in the Federal
Register, followed by a public hearing that provided the public with
the opportunity to comment. The final notice synthesized what was
learned and it described NHTSA's final requirements. A similar process
will be used before new technologies are included in the future.
DOT Faces a Foresight Communication Challenge:
DOT faces the challenge of conveying the potentially difficult issues
posed by high-clockspeed trends and the complexity of forces that could
affect highway fatalities in the mid-term and long-term future.
Additionally, such information would ideally be provided in a timely
fashion for congressional deliberations. Experts we consulted with
emphasized the importance of timing and suggested the need to reexamine
DOT's capability for producing such information.
We have suggested the value of a systematic, organized approach to
guiding the development of information in the area of program
evaluation (INTOSAI 2007). In a December 2007 report, we noted also
that a set of analytical tools can help policymakers transform
government to better meet the demands of the 21st century (GAO 2007).
According to that report, the consistent use of these items will help
policymakers (1) reach consensus on the outcomes Americans most want
their government to achieve, (2) increase transparency and
accountability, (3) better prioritize competing demands, (4) make more
informed decisions, and (5) modernize federal operations and
management. An example of this approach might be adapting GAO's
framework for exercising foresight for communications with the
Congress.
In addition to conceptual frameworks such as those developed for this
report, DOT could consider the value of foresight tools for enhancing
communication to the Congress on fast-paced and complex trends that
require responses by multiple stakeholders. These could include:
Technology and policy roadmaps that describe multiple trends, multiple
governance options, and the responsibilities of multiple parties:
The technology roadmap can be used to illustrate the nature, rate, and
direction of future science and technology developments related to in-
vehicle, vehicle-to-vehicle, and vehicle-road technologies and systems
(FURORE 2003). An expanded version could include attention to
electronic driver distractions.
A policy-oriented version of the roadmap tool illustrates one model
illuminating strategic responses and options for action, as well as
communicating organizational responsibilities. DOT's "Hydrogen Roadmap"
was developed in 2005 as the guiding document for the DOT Hydrogen
Safety Research, Development, Demonstration, and Deployment programs
(DOT 2005). It serves as an outreach document for communication,
coordination, and collaboration with other federal agencies, industry,
the public, and the Congress.[Footnote 87]
Foresight methods that inform strategy development by addressing
multiple trends:
One technique, known as backcasting, generates alternative scenarios
for achieving a desired future goal, such as President Kennedy's goal
of placing a man on the moon. Backcasting can account for sources of
uncertainty and inform deliberations about how to design the most
robust strategy for achieving that goal. Another technique is that of
technology assessment, which provides a formal analysis of the
implications of technology developments and can include an
understanding of the unintended consequences of new technologies.
Using foresight methods within an organizationwide capability:
The foresight program in the United Kingdom's Government Office for
Science uses a model for conducting foresight studies that stresses the
importance of an evidence-based approach and includes attention to
multiple trends. In addition, stakeholders are involved both during and
after a study is conducted. This unit first identifies an emerging
development, such as obesity or intelligent infrastructure systems for
transportation; describes the range of factors influencing this
development and the interaction of such factors; and identifies
opportunities for policy intervention. The unit develops scenarios that
can then be used to explore the potential effect of different response
options. Stakeholders are involved at each stage, from identifying
issues to making action plans. They can include local governments,
various ministries within the U.K. government, research associations,
and industry groups.[Footnote 88]
Transportation and foresight experts, as well as DOT officials, cited
the importance of communicating information on trends and their
implications through the reauthorization of funding for surface
transportation programs, and they highlighted the importance of timely
information for use in this process. Our interviews and review of the
foresight literature indicate the importance of identifying which
future horizon is appropriate for study; integration with strategic
planning and performance measurement processes; developing, testing,
and disseminating new methodological approaches so as to inform both
action and evidence on emerging trends in the future; providing an
ongoing monitoring function; and working with a wide range of
stakeholders.[Footnote 89]
Transportation experts we interviewed, as well as DOT officials, noted
the value of reexamining the potential of existing resources within the
department for ongoing multimodal or goal-specific trend
identification, analysis, and communications. According to DOT, these
resources include RITA's Volpe Center and BTS.[Footnote 90] By better
harnessing such resources, as experts noted, DOT and its
administrations could enhance their ability to understand and
communicate the implications of 21st century trends relevant to highway
safety.
Conclusions:
High-clockspeed technology trends may shape the future of highway
safety, and timely action may be needed to pursue opportunities and
counter threats. In instances where data are somewhat limited, DOT has
sometimes taken action--for example, incorporating information on lane
departure warning and forward-collision avoidance systems on NCAP's new-
car labels and pursuing safety applications for vehicle-road
communications, such as warnings. But while pursuing the potential
opportunity represented by new crash avoidance technologies, DOT
decided that at this time it will not self-initiate (1) empirical
research on countermeasures for electronic driver distractions or (2)
actions to counter this threat. However, DOT has conducted and
sponsored a variety of research studies aimed at understanding the
process of distraction and the impact that electronic distractions may
have on safety. Appropriate decisions are needed for responses to both
opportunities and threats and it is important for decision-making
criteria to be clear and transparent, to be well documented, and--in
the case of new trends for which definitive safety information is
lacking--to address issues of uncertainty. Given the uncertainties that
characterize high-clockspeed trends, a number of relevant approaches
such as anticipatory risk management (which could consider the risks,
for example, of taking no action on a somewhat uncertain but
potentially damaging trend) could help avoid situations in which
highway safety is improved in one area but deteriorates in another.
Additionally, DOT and others emphasize obtaining high-quality evidence
and using this as a basis for decisions, wherever possible. DOT's
established methods of collecting high-quality safety data--as well as
some innovative, newer methods that DOT has used--may not be
sufficiently timely or otherwise suitable for studying the safety
impacts of rapidly changing technology-based trends. It remains unclear
whether DOT's current data collection systems can adequately track new
trends and their impact on highway safety. To enhance safety with
informed decision making and the best possible data (data to better
establish the size of a new safety problem, for example, or the
trajectory of its changing size), new approaches to developing
evidence--such as using wireless technology to automatically collect
data on crashes or DOT-wide guidance on using data in decision making
on highway safety--would help determine when it is appropriate for a
DOT administration to undertake action-oriented research, initiate
consumer education programs, or take various kinds of actions specific
to emerging or evolving trends that affect highway safety.
Finally, DOT's communication with stakeholders, including congressional
policy makers, about future highway safety trends might be more
comprehensive and timely. In particular, by providing comprehensive and
timely information about what is known on trends for which it believes
the data are insufficient, DOT could provide the Congress and others
with potentially important information they could use in determining
how to set national funding and research priorities, especially as the
next process for reauthorizing surface transportation funding
approaches. Consideration of U.S. and international models for
communicating and developing information, as well as frameworks such as
the one we have developed, could help DOT determine how to structure
information for the Congress, even when data on new safety trends are
unclear.
Recommendations for Executive Action:
In order to develop an approach to decision making and the development
of evidence on high-clockspeed trends affecting highway safety that are
characterized by uncertainty, we recommend that the Secretary of
Transportation consider and evaluate practices and principles for
making decisions under conditions of uncertainty and for using data in
such decision making and, on that basis, develop an approach to guide
decision making on high-clockspeed trends that, although somewhat
uncertain, may affect highway safety. We further recommend that the
Secretary of Transportation evaluate whether or not new approaches to
data collection are needed to better track new trends related to
highway safety.
In addition, in order to improve the information available to the
Congress for reauthorization, we recommend that the Secretary of
Transportation (1) analyze and report on trends currently anticipated
to affect highway safety through 2020 and beyond in a systematic
fashion--including information on high-clockspeed trends, discussion of
evidence about these and other individual trends, their implications
and potential interactions, and DOT responses--and (2) determine, in
consultation with relevant congressional committees, schedules for
periodic reporting that will be sufficiently frequent to update the
Congress on fast-changing trends.
Agency Comments and Our Evaluation:
We provided a copy of a draft of this report to the Secretary of
Transportation for review and comment. DOT provided written comments
(see Appendix X). DOT disagreed with our recommendation that the
Secretary develop a new approach to guide decision-making under
conditions of uncertainty. Specifically, DOT states that (1) future
developments are difficult to foresee; (2) the application of a new
decision-making approach that takes account of uncertainty would
require extremely large amounts of data, sophisticated analytic tools,
skilled practitioners, and significant investment of resources--and
would divert substantial resources from ongoing DOT work addressing
proven safety problems; and (3) new approaches to decision-making would
not provide the rigorous results needed for regulatory action.
We continue to recommend that the Secretary of Transportation develop a
new decision-making approach for the following reasons.
* Although future developments are difficult to foresee, the decision-
making approaches we discuss include strategies that avoid
interventions based on a single assumed future course of events. Such
strategies do not require foreseeing specific future developments.
Rather, they recognize uncertainty--based on limited evidence and the
possibility of alternative future developments. For example, in this
report we describe strategies that literature and experts have
suggested, including (1) developing optional plans, such as "Plan A"
and "Plan B," so that either can be used depending on how the future
unfolds; and (2) choosing robust solutions that are expected to work
reasonably well across multiple possible future developments.
* While DOT pointed to a need for large amounts of data, skilled
practitioners, and other requirements, DOT has undertaken foresight
efforts in the past despite these challenges. For example, as discussed
earlier in this report, DOT reported trend analyses in The Changing
Face of Transportation (DOT 2000); additionally, DOT's Volpe Center
recently applied the "technology roadmap" to identify potential safety
issues in the development of new materials for use in automobiles
(Brecher 2007). Further, we believe a new decision-making approach
would not require DOT to shift significant resources away from current
efforts to address established safety problems. Some responses to new
or evolving trends may be relatively low cost. For example, a
relatively low-cost response to an evolving trend that the states are
beginning to address might be a research project that compares levels
of success under different programs that have been adopted in various
states; the results could then be provided to the states, including
those states that are currently considering similar programs.
* We agree that DOT regulatory action requires rigorous evidence and
recognize that developing timely, rigorous evidence on new trends may
be challenging. However, governance tools other than regulatory action
are available for shaping or countering such trends. Conducting
exploratory research on how to effectively shape or counter a trend may
inform future decision-making at the federal or the state level. (See
figures 9, 11, and 13 of this report, which outline a range of
governance tools.)
Additionally, DOT notes that it is addressing the older driver trend,
as described in a recent GAO report. We agree that DOT is making
anticipatory decisions in some cases (and this report describes DOT
efforts to shape the development of crash-avoidance technologies).
However, we continue to recommend that DOT develop a new approach to
guide its decision making on high-clockspeed trends because DOT is
currently addressing some trends that affect highway safety and not
others, without basing these decisions on criteria designed to take
uncertainty in account.
DOT did not specifically comment on our recommendation regarding new
methods for developing more timely evidence on high-clockspeed trends
or our recommendation for improved communications with Congress and
other stakeholders.
DOT disagreed with the wording of our draft conclusion regarding its
response to fast-evolving electronic driver distractions. We reworded
this conclusion to clarify that DOT has researched the process of
driver distraction and the safety impact of electronic distractions,
but is not at this time self-initiating empirical research on
countermeasures (for example, it is not evaluating the effectiveness of
existing countermeasures).
Finally, DOT provided technical comments, which we addressed throughout
this report, as appropriate.
We are sending copies of this report to interested congressional
committees, the Secretary of Transportation, and the Administrators of
NHTSA, RITA, FHWA, and FMCSA. We will also make copies available to
others on request. In addition, the report will be available at no
charge on the GAO Web site at [hyperlink, http://www.gao.gov].
If you or your staff have any questions, please contact Nancy Kingsbury
at (202) 512-2700 or kingsburyn@gao.gov or Katherine Siggerud at (202)
512-2834 or siggerudk@gao.gov. Contact points for our Offices of
Congressional Relations and Public Affairs are on the last page. GAO
staff who made contributions to this report are listed in appendix XI.
Sincerely yours,
Signed by:
Nancy R. Kingsbury:
Managing Director, Applied Research and Methods:
Signed by:
Katherine A. Siggerud:
Managing Director, Physical Infrastructure Issues:
[End of section]
Appendix I: Exercising Foresight: Three Agency Activities and Related
Challenges:
Addressing high-clockspeed trends with the purpose of aiding future
progress toward basic goals--across a time horizon extending more than
5 years forward--is one way that an agency can exercise foresight.
[Footnote 91] Three agency activities are relevant:
* Deciding how to respond to trends that can affect important future
outcomes. Statements of the Comptroller General of the United States
and foresight literature have urged that current decisions or near-term
actions be taken with the purpose of shaping future outcomes (see
Walker 2005; Dewar 2006; Rejeski 2003). According to risk management
experts and strategic planning literature, goal-achievement efforts
should include responses to both opportunities and threats (Harvard
Business School 2005, ch. 1).
* Developing additional evidence on such trends. The Comptroller
General stated that an evidence-based approach to planning for the
future is essential to helping policy makers expand their time horizon
(Walker 2007). The evidence-based policy perspective emphasizes the
development of high-quality evidence and its use in decision making.
* Communicating with the Congress and others about trends and
strategies to address them. Our recent report on stewardship and 21st
challenges emphasizes the need for agencies to communicate with the
Congress about developments with implications for the future in a way
that is transparent, timely, and useful (GAO 2007).
Foresight literature and experts point to a dynamic process in which
earlier decisions about how to respond to a new trend are revised as
uncertainty about that trend is reduced--for example, through the
development of new evidence on the trend's impact (Younes 2005).
Therefore, we view the three activities as part of an ongoing iterative
process.
Additionally, literature and experts recognize that as an agency
addresses high-clockspeed trends through the activities outlined above,
it may encounter challenges, including:
* A challenge of decision making under conditions of uncertainty.
Uncertainty may characterize various aspects of high-clockspeed or
other trends or their future directions (Courtney, Kirkland, and
Viguerie 1997; Popper, Lempert, and Bankes 2005).
* A challenge of developing evidence when existing data systems or
analysis methods are not relevant. Statistical experts told us that
data systems and methods that may have worked well in the past may not
be suited to studying high-clockspeed or other new trends.
* A challenge of providing policy-relevant information about the
future, such as potential unintended consequences. The foresight
literature highlights the difficulty and importance of identifying such
consequences for high-clockspeed trends because there may be little
time to anticipate or react to them (Rejeski 2003, 56-57).
[End of section]
Appendix II: Private Sector, University, and Other Experts We
Interviewed:
In addition to interviewing DOT and TRB officials and staff, and
officials from some states, we interviewed the experts listed in this
appendix.[Footnote 92]
Area: Automotive and highway safety;
Expert: Thomas A. Dingus;
Organization: Virginia Tech Transportation Institute, Blacksburg, Va.
Area: Automotive and highway safety;
Expert: Clarence M. Ditlow;
Organization: Center for Auto Safety, Washington, D.C. .
Area: Automotive and highway safety;
Expert: Gerald A. Donaldson, Jacqueline S. Gillan, and Henry M. Jasney;
Organization: Advocates for Highway Safety, Washington, D.C.
Area: Automotive and highway safety;
Expert: Mortimer L. Downey III;
Organization: PB Consult, Washington, D.C.
Area: Automotive and highway safety;
Expert: James C. Fell;
Organization: Pacific Institute for Research and Evaluation, Calverton,
Md.
Area: Automotive and highway safety;
Expert: Barry Felrice and David Henry;
Organization: Chrysler LLC, Auburn Hills, Mich.
Area: Automotive and highway safety;
Expert: Michael D. Freitas;
Organization: Ygomi LLC, Washington, D.C.
Area: Automotive and highway safety;
Expert: Paul M. Horn (retired);
Organization: IBM, Armonk, N.Y.
Area: Automotive and highway safety;
Expert: Anthony R. Kane and Ken F. Kobetsky;
Organization: American Association of State Highway and Transportation
Officials, Washington, D.C.
Area: Automotive and highway safety;
Expert: Thomas M. Kowalick;
Organization: Click Incorporated, Southern Pines, N.C.
Area: Automotive and highway safety;
Expert: Robert C. Lange with Stephen G. Gehring and Stephen E. O'Toole;
Organization: General Motors Corporation[A].
Area: Automotive and highway safety;
Expert: Adrian Lund, Anne T. McCartt, and Richard A. Retting;
Organization: Insurance Institute for Highway Safety, Arlington, Va.
Area: Automotive and highway safety;
Expert: Robert E. Martinez;
Organization: Norfolk Southern Corp., Norfolk, Va.
Area: Automotive and highway safety;
Expert: Carl E. Nash;
Organization: National Crash Analysis Center, George Washington
University, Washington, D.C.
Area: Automotive and highway safety;
Expert: Priya Prasad;
Organization: Ford Motor Company, Dearborn, Mich.
Area: Automotive and highway safety;
Expert: George L. Reagle;
Organization: George Reagle and Associates, Columbia, Md.
Area: Electronics, communications;
Expert: Michael F. Altschul and Robert F. Roche;
Organization: CTIA--The Wireless Association,® Washington, D.C.
Area: Electronics, communications;
Expert: William L. Ball;
Organization: OnStar™, General Motors Corporation, Detroit, Mich.
Area: Electronics, communications;
Expert: Joseph Brennan;
Organization: Trinity-Noble LLC, Doylestown, Pa.
Area: Electronics, communications;
Expert: Charles L. Eger;
Organization: Motorola Inc., Washington, D.C.
Area: Electronics, communications;
Expert: Brian F. Fontes;
Organization: AT&T, Washington, D.C.
Area: Electronics, communications;
Expert: Arlene Harris;
Organization: GreatCall Inc., Del Mar, Calif.
Area: Electronics, communications;
Expert: Steve Koenig , J. David Grossman, and Ellen Savage;
Organization: Consumer Electronics Association, Arlington, Va.
Area: Electronics, communications;
Expert: Tyler C. Messa;
Organization: Telecommunications Industry Association, Arlington, Va.
Area: Electronics, communications;
Expert: Paul Nash and Charon Phillips;
Organization: Verizon Wireless, Washington, D.C.
Area: Electronics, communications;
Expert: T. Russell Shields, Richard Weiland, and Sheryl J. Wilkerson;
Organization: Ygomi LLC, Oak Brook, Ill., and Washington, D.C.
Area: Public policy, research methods, and foresight;
Expert: John C. Bailar III;
Organization: National Academy of Sciences, Washington, D.C., and the
University of Chicago, Chicago, Ill.
Area: Public policy, research methods, and foresight;
Expert: Joseph Coates;
Organization: Consulting Futurist Inc., Washington, D.C.
Area: Public policy, research methods, and foresight;
Expert: Kenneth W. Hunter;
Organization: Institute for Global Chinese Affairs, University of
Maryland, College Park.
Area: Public policy, research methods, and foresight;
Expert: Mary Grace Kovar;
Organization: Consultant, Washington, D.C.
Area: Public policy, research methods, and foresight;
Expert: Rose Marie Martinez;
Organization: Institute of Medicine, National Academies, Washington,
D.C.
Area: Public policy, research methods, and foresight;
Expert: Paul Posner;
Organization: George Mason University, Fairfax, Va.
Area: Public policy, research methods, and foresight;
Expert: David W. Rejeski;
Organization: Woodrow Wilson International Center for Scholars,
Washington, D.C.
Area: Public policy, research methods, and foresight;
Expert: Fritz J. Scheuren;
Organization: National Opinion Research Center, University of Chicago,
Chicago, Ill.
Source: GAO.
Note: For experts with expertise in more than one category, we selected
the category that seemed most relevant.
[A] Robert Lange is in General Motors' Warren, Michigan, office.
Stephen O'Toole's and Stephen Gehring's offices are in Washington, D.C.
[End of table]
We met with many automobile manufacturers' and suppliers'
representatives in meetings the Alliance of Automobile Manufacturers
and Association of International Automobile Manufactures (AIAM)
convened.[Footnote 93]
Finally, we obtained background information on foresight from a number
of experts whose work we cite in the report. Others who provided
significant background information include Clement Bezold, Institute
for Alternative Futures, Alexandria, Va.; Peter C. Bishop, University
of Houston, Houston, Texas; Theodore J. Gordon, Millennium Project,
World Federation of United Nations Associations, Washington, D.C.;
Donna M. Heivilin, formerly with Applied Research and Methods, U.S.
Government Accountability Office; Andy Hines, Social Technologies Group
Inc., Washington, D.C.; Wendy Schultz of Infinite Futures, Oxford,
England; and Edie Weiner of Weiner, Edrich, Brown Inc., New York, N.Y.
[End of section]
Appendix III: Review of Studies of Driver Phoning and Highway Safety:
For our review of driver phoning and safety risks, we searched (1) all
peer-reviewed journals in ProQuest and EconLit and (2) articles listed
for the journal of the Transportation Research Board, a peer-reviewed
journal in the TRIS database, through October 2007. We also considered
the 100-car study and a literature review from the Insurance Institute
of Highway Safety. This yielded:
* 13 primary studies of driver phoning and crash risks of diverse
designs (table 4 summarizes the results, and table 5 gives quality
descriptors),[Footnote 94]
* other primary studies of impacts on driver performance (for example,
braking time) without indicating crash risks, and:
* three review articles that focused on driver phoning.
Table 4: Impact of Driver Phoning on Crash Risk: 13 Primary Studies:
Effect of driver phoning: Increase in crash risk;
Number of primary studies and types of design: 12 of 13 studies.
Effect of driver phoning: Estimated size of crash risk;
Number of primary studies and types of design: 10 of 12 studies
estimating crash risk.
Effect of driver phoning: Quadrupled or quintupled overall risk;
Number of primary studies and types of design: 4 studies (2 "real
world" case-control with record checks conducted in Canada and
Australia; 2 U.S. simulator studies)[A].
Effect of driver phoning: Doubled or tripled overall risk;
Number of primary studies and types of design: 3 studies (2 U.S.
simulator studies; 1 U.S. self-report survey)[B].
Effect of driver phoning: Doubled risk under one condition but limited
or no impact for another condition;
Number of primary studies and types of design: 2 studies (in one, the
100-car study, risk more than doubled when dialing a call; in the
other, a study of decisions-made-on-a-track in Canada, risk doubled on
wet pavement).
Effect of driver phoning: Overall limited or small impact;
Number of primary studies and types of design: 1 study (self-report
survey in Brazil; statistically significant 12% increase)[B].
Effect of driver phoning: Amount of increase not specified[C];
Number of primary studies and types of design: 2 studies (1 U.S.
simulator study; 1 U.S. self-report survey).
Effect of driver phoning: No increase in risk;
Number of primary studies and types of design: 1 study (U.S. simulator
study).
Effect of driver phoning: Total sampling;
Number of primary studies and types of design: 13 studies (6 simulator
studies, 2 "real world" case-control with record checks, 3 self-report
surveys, 1 "decisions on a track" study, and the 100-car study).
Source: GAO analysis of peer-reviewed literature and the 100-car study.
Note: These studies are listed by design category toward the end of
this appendix.
[A] One of the simulator studies included regular drivers and airplane
pilots, reporting results separately for the two groups. For this
table, we considered only results reported for the regular drivers.
[B] The survey documented an association between self-reports of amount
of cell phone use while driving and crashes; it excluded reports of
crashes occurring while the driver was using the phone.
[C] These studies reported (1) an association between phoning and
crashes or (2) an increased risk with amount not specified.
[End of table]
Table 5: Technical Adequacy Descriptors: Primary Studies of Crash Risk
by Design Category:
Technical adequacy descriptor: Validity: Controlled comparison[B];
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": Yes, strength;
Design category: Retrospective: 2 case-control studies with record
checks: Limited[C];
Design category: Retrospective: 3 self-report surveys: Limited[C];
Design category: Mixed[A]: The 100-car study: Limited[C].
Technical adequacy descriptor: Validity: Accurate measures of phoning:
e.g., direct observation;
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": Yes, strength;
Design category: Retrospective: 2 case-control studies with record
checks: Mixed[D];
Design category: Retrospective: 3 self-report surveys: Limited;
Design category: Mixed[A]: The 100-car study: Yes, strength.
Technical adequacy descriptor: Validity: "Real world" conversations,
driving, and crashes;
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": No, weakness[E];
Design category: Retrospective: 2 case-control studies with record
checks: Yes, strength;
Design category: Retrospective: 3 self-report surveys: Yes, strength;
Design category: Mixed[A]: The 100-car study: Mixed[F].
Technical adequacy descriptor: Validity: Free of "reactive effects":
e.g., subjects' behavior or statements not biased by knowing they were
being studied;
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": No, weakness[G];
Design category: Retrospective: 2 case-control studies with record
checks: Yes, strength;
Design category: Retrospective: 3 self-report surveys: No, weakness;
Design category: Mixed[A]: The 100-car study: Possibly[H].
Technical adequacy descriptor: Reliability: reproducibility and
consistency;
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": Mixed: small samples[I] but negative impact
observed for 6 of the 7 studies;
Design category: Retrospective: 2 case-control studies with record
checks: Yes, strength: large samples, consistent results for both
studies;
Design category: Retrospective: 3 self-report surveys: Yes, strength:
large samples, all 3 found some association of phoning with crashes;
Design category: Mixed[A]: The 100-car study: No, weakness: only 100
cars.
Technical adequacy descriptor: Generalizability: to the full U.S.
population of drivers, crashes;
Design category: Prospective or controlled: 7 driving simulator studies
and "decisions on a track": No, weakness: volunteers from varied
population groups;
Design category: Retrospective: 2 case-control studies with record
checks: Possibly generalizable but conducted in Canada and Australia;
Design category: Retrospective: 3 self-report surveys: No, weakness: an
Internet survey, a college classroom survey, and an intercept survey
(conducted in Brazil);
Design category: Mixed[A]: The 100-car study: No, weakness: limited to
one geographic area, omitted unlicensed drivers.
Source: GAO analysis, peer-reviewed literature, and the 100-car study.
Note: These studies are listed by design category toward the end of
this appendix.
[A] Mixed design means that the study had elements of prospective and
retrospective designs; the 100-car study was prospective in that it set
up observations in advance (and subjects knew they were being studied),
but driver behavior was not controlled and the analysis was
retrospective.
[B] Random assignment, or each subject served as his or her own
control.
[C] Retrospective studies, including case-control, surveys, and the 100-
car study, do not involve random assignment to comparison groups (such
as phoning while driving versus driving only) or other prospective
controls to ensure balanced, unbiased comparisons. Statistical controls
may be applied but these adjustments are designed to reduce imbalances
and do not provide as strong an assurance of equivalence as prospective
controls.
[D] These studies checked cell phone records and compared recorded
times of crashes and calls, but it has been argued that some calls made
after a crash could have been misidentified as precrash calls (because,
for example, the recorded time of the crash may not have been exact).
[E] One of the simulator studies featured "naturalistic" conversations;
that is, a research assistant talked with subjects in a way intended to
be similar to a real-world conversation.
[F] In the 100-car study, the conversations and the driving were real-
world, but near misses were the primary basis for estimating crash
risks.
[G] In one simulator study, subjects were told they were testing
software for the driving simulator. Thus, although they knew they were
being studied, they did not know that the purpose was to evaluate their
driving skills.
[H] Drivers in the 100-car study were aware that video cameras and
recording devices were installed in the vehicles they were driving,
suggesting the potential for reactive effects; however, over time,
some--perhaps many--drivers may have become used to these devices and
disregarded them.
[I] Range of sample sizes for these 7 studies: 20 to 55 subjects. (One
simulator study had 55 regular drivers and 56 airplane pilots; we
considered only the results for regular drivers.)
[End of table]
As table 4 shows, 12 of the 13 primary studies indicated at least some
increase in risk as a result of driver phoning. Two indicated an
increase in risk without estimating the size of the increase. Of the
remaining 10 studies (all estimated the size of the increase in risk),
one reported an overall small (12 percent) increase, two reported that
under some circumstances (that is, dialing a call, driving on wet
pavement) there was a doubling of risk or higher impact, and the
remaining 7 reported a twofold to fivefold overall increase.
Of the 13 primary studies, 5 compared safety risks for handheld and
hands-free phoning.[Footnote 95] None found any difference. (The 13
primary studies of crash risk and the three reviews of driver phoning
studies are listed at the end of this appendix.)
A key feature of the 100-car study was its ability to compare the
percentages of crashes and near misses associated with various
activities engaged in by participating drivers.[Footnote 96] Results
may not be generalizable but are suggestive. Taken together, dialing
and talking on or listening to a handheld device were associated with
over 7 percent of observed crashes and near misses. Eating while
driving was associated with slightly over 2 percent; applying makeup,
less than 2 percent; reaching for a moving object, 1 percent; and
drinking from an open container, less than half of 1 percent (Klauer
and others 2006, 33).[Footnote 97]
Other results concerning driver performance are included in some of the
studies of crash risk and in other peer-reviewed primary studies that
did not examine crash risk. These other results indicate that driving
performance is degraded by driver phoning. For example, drivers using a
cell phone braked more slowly in response to a lead vehicle's braking.
Each of the three review articles concluded that driver phoning
increased crash risk or degraded driving performance:
* "there is a growing body of evidence...that cell phone use
substantially increases crash risk" and that risk is not eliminated by
driver use of a "hands-free" phone (McCartt, Helinga, and Bratiman
2006, 102);
* there are "clear costs to driving performance [primarily in terms of
reaction time] when drivers [are] engaged in cell phone conversations"
(Horrey and Wickens 2006, 196); and:
* "using a mobile phone when driving...disturbs driving through a
diminished field of attention, longer detection times to, e.g., changes
in dynamic traffic conditions, longer braking reaction-times...and
greater lateral deviations on the road...[and] complex conversations
disturb more than simple conversations" (Svenson and Patten 2005, 195).
Studies of Driver Cell Phone Use:
Primary Studies of Crash Risk Identified in GAO Searches by Study
Design:
Driving Simulator Studies and "Decisions on a Track":
Abdel-Aty, Mohamed. 2003. Investigating the Relationship between
Cellular Phone Use and Traffic Safety. Institute of Transportation
Engineers. ITE Journal 73, no. 10 (October):38-42.
Cooper, Peter J., and Yvonne Zheng. 2002. Turning Gap Acceptance
Decision-Making: The Impact of Driver Distraction. Journal of Safety
Research 33, no. 3 (Fall): 321-35.
Hunton, James, and Jacob M. Rose. 2005. Cellular Telephones and Driving
Performance: The Effects of Attentional Demands on Motor Vehicle Crash
Risk. Risk Analysis (Oxford) 25, no. 4 (August): 855-66.
Rakauskas, Michael E., Leo J. Gugerty, and Nicholas J. Ward. 2004.
Effects of Naturalistic Cell Phone Conversations on Driving
Performance. Journal of Safety Research 35, no. 4:453-64.
Schattler, Kerrie L., and others. 2006. Assessing Driver Distractions
from Cell Phone Use While Driving: A Simulator-Based Study. Paper
submitted at the 85th Annual Meeting of the Transportation Research
Board, Washington, D.C.: January.
Strayer, David L., and Frank Drews. 2004. Profiles in Driver
Distraction: Effects of Cell Phone Conversations on Younger and Older
Drivers. Human Factors 46, no. 4 (Winter):640-49.
Strayer, David L., Frank A. Drews, and Dennis Crouch. 2006. A
Comparison of the Cell Phone Driver and the Drunk Driver. Human Factors
48, no. 2 (Summer):381-91.
Case Control Studies with Record Checks:
McEvoy, Suzanne, and others. 2005. Role of Mobile Phones in Motor
Vehicle Crashes Resulting in Hospital Attendance: A Case-Crossover
Study. British Medical Journal (International Edition: London) 331, no.
7514 (August 20-27):428-30.
Redelmeier, Donald A., and Robert J. Tibshirani. 1997. Association
between Cellular-Telephone Calls and Motor Vehicle Collisions. The New
England Journal of Medicine 336, no. 7 (February 13): 453-58.
Self-Report Surveys:
Hahn, Robert W., and James E. Prieger. 2006. The Impact of Driver Cell
Phone Use on Accidents. B. E. Journals in Economic Analysis and Policy:
Advances in Economic Analysis and Policy 6, no. 1:1-37.
Paulo, Loureiro, Adolfo Sachsida, and Tito Moreira. 2004. Traffic
Accidents: An Econometric Investigation. Economics Bulletin 18, no. 3:
1-7.
Seo, Dong-Chul, and Mohammad R. Torabi. 2004. The Impact of In-Vehicle
Cell-Phone Use on Accidents or Near-Accidents among College Students.
Journal of American College Health 53, no.3 (November/December): 101-
07.
The 100-Car Study:
Klauer, S. G., and others. 2006. The Impact of Driver Inattention on
Near-Crash/Crash Risk: An Analysis Using the 100-Car Naturalistic
Driving Study Data. Washington, D.C.: U.S. Department of
Transportation, April. DOT HS 810 594,
VTRC (Virginia Transportation Research Council). 2005. VTRC a Co-
sponsor of Groundbreaking Study of Driver Behavior. Press release,
August 15. Accessible at [hyperlink,
http://vtrc.virginiadot.org/BriefDetails.aspx?Id=19].
Two Review Articles Identified in GAO Searches:
Horrey, William J., and Christopher D. Wickens. 2006. Examining the
Impact of Cell Phone Conversations on Driving Using Meta-Analytic
Techniques. Human Factors 48, no.1 (Spring):196-205.
Svenson, Ola, and Christopher J. D. Patten. 2005. Mobile Phones and
Driving: A Review of Contemporary Research. Cognitive Technology
Journal 7:182-97.
Review Article the Insurance Institute of Highway Safety Provided to
Us:
McCartt, Anne T., Laurie Helinga, and Kel Bratiman. 2006. Cell Phones
and Driving: Review of Research. Traffic Injury Prevention 7:89-106.
Definition of Technical Adequacy:
In this report, we consider technical adequacy to be composed of
validity, reliability, and generalizability:
* Validity is defined here as unbiased counts or measurement of what
one intends to measure--for example, safety outcomes measured by actual
crashes or road fatalities rather than by surrogate measures such as
near misses (unless the surrogate measures have been shown to be highly
correlated with actual crashes or fatalities in a relevant research
context).
* Reliability is defined here as reproducibility and consistency, or
the absence of random error--that is, similar results would be obtained
if the study were repeated using the same procedures. In studies in
which driver reactions may vary, a key factor contributing to
reliability is a sufficiently large study, and a study with few drivers
would be likely to yield less reliable results than a study with a much
larger sample of drivers.
* Generalizability is defined here as the applicability of study
results to the population of interest--for example, a study that draws
a representative sample from all drivers on U.S. roads would be
generalizable in that it could be expected to produce results
characterizing that population as a whole.
[End of section]
Appendix IV: Trends for Vulnerable Road-User Groups:
Trends for certain vulnerable road-user groups point to heightened
safety challenges in the years ahead, and DOT is taking some steps to
address these.
* By 2025, the annual number of road fatalities for older drivers may
be double what it was in 2005. The main reason for the projected
increase is that the first members of the baby boom generation will
reach their 65th birthday in 2011, and the number and percentage of
Americans older than 65 will steadily increase for several years. In
response to this trend, DOT is examining various issues relevant to
older drivers (see Band and Perel 2007).
* Motorcycle riders are vulnerable in any type of crash, and the
numbers of riders and fatalities have risen in recent years. DOT has
anticipated more motorcycles on U.S. roads and has begun taking steps
to address related issues--by, for example, examining a broad range of
alternative options for reducing motorcycle fatalities (DOT 2007). Most
recently, some have speculated that higher than expected increases in
motorcycle VMT (or increased scooter use) will occur if fuel prices
continue to rise.[Footnote 98]
* Occupants of light passenger vehicles in crashes with heavy trucks
represent another vulnerable group, and this vulnerability may
increase. A recent projection from the Energy Information
Administration has suggested that by 2020, VMT for freight trucks
weighing over 10,000 pounds will increase by about 30 percent relative
to mileage observed in 2006, although the changing economy and rising
costs of fuel may limit anticipated increases.
* Occupants of smaller cars are generally vulnerable in crashes.
[Footnote 99] The category with the smallest compact cars or minis
(such as the Scion and Aveo) represents a very small percentage of the
fleet but may have been growing in recent years, as seen in figure 17.
The figure also shows that sales of other small cars recently spiked,
possibly as a result of rising fuel prices. At the same time, large
sport utility vehicles appear to be declining in popularity.
Figure 17: Small Cars as a Percentage of Passenger Vehicle Sales,
January 2003 to June 2008:
[Refer to PDF for image]
This figure is a multiple line graph depicting the following data in
percentage of all passenger vehicle sales in these three categories:
1) All small cars (includes all cars in Ward Automotive Group’s "lower
small car,” "upper small car," and "small specialty car" categories.
Ward's defines these categories primarily based on a length of less
than 180 inches, but includes other criteria as well, such as height
and volume).
2) Upper small cars (includes all cars in Ward Automotive Group’s
“upper small car” category and those cars in the "small specialty car"
category between 170 and 180 inches long. Examples are such cars as
Ford Focus, Nissan Sentra, Toyota Corolla, Volkswagen Jetta, and
Volkswagen Rabbit).
3) Lower small cars (includes all cars in Ward’s “lower small car”
category and cars in the “small specialty car” category less than 170
inches long. Examples are such cars as the Chevrolet Aveo, Honda Fit,
Scion xA, Mini Cooper, and Volkswagen Beetle).
Month and year: January 2003;
All small cars: 15;
Upper small cars: 13.5;
Lower small cars: 1.5.
Month and year: February 2003;
All small cars: 14.5;
Upper small cars: 13.2;
Lower small cars: 1.4.
Month and year: March 2003;
All small cars: 15.2;
Upper small cars: 13.8;
Lower small cars: 1.4.
Month and year: April 2003;
All small cars: 13.6;
Upper small cars: 12.2;
Lower small cars: 1.3;
Month and year: May 2003;
All small cars: 14.2;
Upper small cars: 12.9;
Lower small cars: 1.3.
Month and year: June 2003;
All small cars: 14.7;
Upper small cars: 13.4;
Lower small cars: 1.3.
Month and year: July 2003;
All small cars: 14.2;
Upper small cars: 13;
Lower small cars: 1.2.
Month and year: August 2003;
All small cars: 13.9;
Upper small cars: 12.7;
Lower small cars: 1.2.
Month and year: September 2003;
All small cars: 13.6;
Upper small cars: 12.4;
Lower small cars: 1.2.
Month and year: October 2003;
All small cars: 13.1;
Upper small cars: 11.9;
Lower small cars: 1.2.
Month and year: November 2003;
All small cars: 13.2;
Upper small cars: 11.5;
Lower small cars: 1.7.
Month and year: December 2003;
All small cars: 11.5;
Upper small cars: 10.3;
Lower small cars: 1.2.
Month and year: January 2004;
All small cars: 12.5;
Upper small cars: 11.5;
Lower small cars: 1.1.
Month and year: February 2004;
All small cars: 13.9;
Upper small cars: 12.7;
Lower small cars: 1.2.
Month and year: March 2004;
All small cars: 13.3;
Upper small cars: 12;
Lower small cars: 1.3.
Month and year: April 2004;
All small cars: 13.6;
Upper small cars: 12;
Lower small cars: 1.6.
Month and year: May 2004;
All small cars: 14.7;
Upper small cars: 13.2;
Lower small cars: 1.5.
Month and year: June 2004;
All small cars: 14.5;
Upper small cars: 12.8;
Lower small cars: 1.7.
Month and year: July 2004;
All small cars: 13.9;
Upper small cars: 12.5;
Lower small cars: 1.4.
Month and year: August 2004;
All small cars: 13.7;
Upper small cars: 12.3;
Lower small cars: 1.5.
Month and year: September 2004;
All small cars: 12.5;
Upper small cars: 11.2;
Lower small cars: 1.3.
Month and year: October 2004;
All small cars: 12.5;
Upper small cars: 11.3;
Lower small cars: 1.3.
Month and year: November 2004;
All small cars: 12.7;
Upper small cars: 11.1;
Lower small cars: 1.5.
Month and year: December 2004;
All small cars: 11.4;
Upper small cars: 10.2;
Lower small cars: 1.2.
Month and year: January 2005;
All small cars: 13.6;
Upper small cars: 12.1;
Lower small cars: 1.5.
Month and year: February 2005;
All small cars: 13.2;
Upper small cars: 12;
Lower small cars: 1.2.
Month and year: March 2005;
All small cars: 13.4;
Upper small cars: 12.2;
Lower small cars: 1.3.
Month and year: April 2005;
All small cars: 15.2;
Upper small cars: 13.7;
Lower small cars: 1.5.
Month and year: May 2005;
All small cars: 14.7;
Upper small cars: 13.1;
Lower small cars: 1.6.
Month and year: June 2005;
All small cars: 13.3;
Upper small cars: 11.8;
Lower small cars: 1.6.
Month and year: July 2005;
All small cars: 13.2;
Upper small cars: 12;
Lower small cars: 1.2.
Month and year: August 2005;
All small cars: 15;
Upper small cars: 13.7;
Lower small cars: 1.3.
Month and year: September 2005;
All small cars: 15.7;
Upper small cars: 14.1;
Lower small cars: 1.6.
Month and year: October 2005;
All small cars: 14.2;
Upper small cars: 12.5;
Lower small cars: 1.7.
Month and year: November 2005;
All small cars: 12.6;
Upper small cars: 11.4;
Lower small cars: 1.2.
Month and year: December 2005;
All small cars: 11.3;
Upper small cars: 10.2;
Lower small cars: 1.1.
Month and year: January 2006;
All small cars: 13.8;
Upper small cars: 12.5;
Lower small cars: 1.2.
Month and year: February 2006;
All small cars: 12.7;
Upper small cars: 11.5;
Lower small cars: 1.2.
Month and year: March 2006;
All small cars: 13.4;
Upper small cars: 12;
Lower small cars: 1.4.
Month and year: April 2006;
All small cars: 16;
Upper small cars: 13.9;
Lower small cars: 2.1.
Month and year: May 2006;
All small cars: 16.6;
Upper small cars: 14.2;
Lower small cars: 2.4.
Month and year: June 2006;
All small cars: 15.9;
Upper small cars: 13.5;
Lower small cars: 2.4.
Month and year: July 2006;
All small cars: 16.6;
Upper small cars: 13.9;
Lower small cars: 2.7.
Month and year: August 2006;
All small cars: 16.4;
Upper small cars: 13.6;
Lower small cars: 2.7.
Month and year: September 2006;
All small cars: 15.1;
Upper small cars: 12.6;
Lower small cars: 2.4.
Month and year: October 2006;
All small cars: 13.3;
Upper small cars: 11;
Lower small cars: 2.3.
Month and year: November 2006;
All small cars: 13.5;
Upper small cars: 11.2;
Lower small cars: 2.3.
Month and year: December 2006;
All small cars: 13.1;
Upper small cars: 11.1;
Lower small cars: 2.
Month and year: January 2007;
All small cars: 14.3;
Upper small cars: 11.9;
Lower small cars: 2.4.
Month and year: February 2007;
All small cars: 13.7;
Upper small cars: 11.4;
Lower small cars: 2.3.
Month and year: March 2007;
All small cars: 14.8;
Upper small cars: 12.2;
Lower small cars: 2.6.
Month and year: April 2007;
All small cars: 15.3;
Upper small cars: 12.7;
Lower small cars: 2.6.
Month and year: May 2007;
All small cars: 17.9;
Upper small cars: 14.9;
Lower small cars: 3.1.
Month and year: June 2007;
All small cars: 17.3;
Upper small cars: 14.3;
Lower small cars: 3.
Month and year: July 2007;
All small cars: 16.3;
Upper small cars: 13.1;
Lower small cars: 3.3.
Month and year: August 2007;
All small cars: 15.6;
Upper small cars: 12.6;
Lower small cars: 2.9.
Month and year: September 2007;
All small cars: 15.3;
Upper small cars: 12.6;
Lower small cars: 2.7.
Month and year: October 2007;
All small cars: 15;
Upper small cars: 12.2;
Lower small cars: 2.8.
Month and year: November 2007;
All small cars: 14.8;
Upper small cars: 12.1;
Lower small cars: 2.7.
Month and year: December 2007;
All small cars: 14;
Upper small cars: 11.5;
Lower small cars: 2.5.
Month and year: January 2008;
All small cars: 15.9;
Upper small cars: 12.8;
Lower small cars: 3.1.
Month and year: February 2008;
All small cars: 15.5;
Upper small cars: 12.4;
Lower small cars: 3.
Month and year: March 2008;
All small cars: 17.1;
Upper small cars: 13.6;
Lower small cars: 3.4.
Month and year: April 2008;
All small cars: 19.3;
Upper small cars: 15.5;
Lower small cars: 3.8.
Month and year: May 2008;
All small cars: 24.3;
Upper small cars: 19.8;
Lower small cars: 4.6.
Month and year: June 2008;
All small cars: 22.6;
Upper small cars: 18.2;
Lower small cars: 4.4.
[End of section]
Appendix V: Illustrations of In-Vehicle Screens for Safety:
Examples of in-vehicle screens for safety purposes are shown in figures
18 and 19.
Figure 18: Photograph: Screen with V2V Icon Warning of a Stopped
Vehicle in the Road Ahead:
[Refer to PDF for image]
Source: GAO.
[End of figure]
Figure 19: Photograph: Backup Camera Screen (Activated When Car Is in
Reverse):
[Refer to PDF for image]
Source: GAO.
[End of figure]
[End of section]
Appendix VI: Information on Selected NCSA Datasets:
Selected datasets maintained by the National Center for Statistics and
Analysis, in the National Highway Safety Administration, are outlined
in table 6, based on information DOT provided us.
Table 6: Selected NCSA Datasets:
Feature: Continuity;
Behavior dataset: NOPUS[A]: Ongoing: annual since 1998;
Behavior dataset: MVOSS[B]: Ongoing: conducted in 1994, 1996, 2001,
2003, and 2007;
Crash dataset: FARS[C]: Ongoing: annual since 1975;
Crash dataset: NASS CDS[D]: Ongoing: annual since 1979;
Crash dataset: NMVCCS[E]: One-time study to be reported in 2008.
Feature: Coverage;
Behavior dataset: NOPUS[A]: For observations of cell phone use: drivers
of passenger vehicles only;
Behavior dataset: MVOSS[B]: Adults 16 and older living in households
with landline telephone;
Crash dataset: FARS[C]: Fatal crashes on roadway;
Crash dataset: NASS CDS[D]: Nationally representative sample of motor
vehicle traffic crashes (1) reported to police and (2) with light
passenger vehicles towed;
Crash dataset: NMVCCS[E]: Nationally representative sample of motor
vehicle crashes (1) reported to police, (2) with light passenger
vehicles towed, (3) one vehicle on scene at time of researcher arrival,
and (4) with emergency medical response.
Feature: Notable omissions;
Behavior dataset: NOPUS[A]: Pedestrians and bicyclists omitted; for
cell phone use, motorcyclists also omitted;
Behavior dataset: MVOSS[B]: Persons in cell-phone-only households or in
households without a phone are omitted;
Crash dataset: FARS[C]: Suicides; off road crashes omitted;
Crash dataset: NASS CDS[D]: Pedestrian, motorcycle, and large truck
crashes without towed passenger vehicle omitted;
Crash dataset: NMVCCS[E]: Similar to omissions noted for NASS CDS. Also
crashes without emergency medical response, crashes with no police
accident report filed, etc., are omitted.
Feature: Approximate sample size;
Behavior dataset: NOPUS[A]: 43,000 drivers observed (annual sample);
Behavior dataset: MVOSS[B]: 12,000 persons interviewed (6,000 per
questionnaire)[F];
Crash dataset: FARS[C]: 39,000 crashes and 43,000 fatalities (annual
census);
Crash dataset: NASS CDS[D]: 4,500 crashes (annual sample);
Crash dataset: NMVCCS[E]: 3,000 crashes (sample for study of crashes in
2005-2006).
Feature: Data sources;
Behavior dataset: NOPUS[A]: Trained observers at selected intersections
and off-ramps;
Behavior dataset: MVOSS[B]: Telephone interviews using structured
questionnaire (landline phones);
Crash dataset: FARS[C]: Police accident reports, vehicle identification
number, and other documents (mostly state-level) such as driver license
information[G];
Crash dataset: NASS CDS[D]: Police accident reports, vehicle
identification number[G,H]; field inspection focused on
crashworthiness[I]; event data recorder[K];
Crash dataset: NMVCCS[E]: Police accident reports, vehicle
identification number[G,H;] field inspection focused on crash
causation[J]; event data recorder[K].
Source: Adapted from information provided by NHTSA's National Center
for Statistics and Analysis.
[A] NOPUS refers to the National Occupant Protection Use Survey,
developed to track the use of seat belts, motorcycle helmets, and child
restraints.
[B] MVOSS refers to the Motor Vehicle Occupant Safety Survey, developed
to obtain a periodic status report of attitudes, knowledge, and self-
reported behavior in areas of motor vehicle occupant protection (seat
belts, child restraints, air bags, crash injury, and emergency medical
response).
[C] FARS refers to the Fatality Analysis Reporting System.
[D] NASS CDS refers to the National Accident Sampling System and the
Crashworthiness Data System. The approximately 4,500 crashes in the CDS
are selected from a nationally representative probability sample for
which police accident reports are collected from the jurisdiction.
[E] NMVCCS (National Motor Vehicle Crash Causation Study) was
authorized by SAFETEA-LU in 2003; the first report is planned for
release in 2008. The approximately 5,000 crashes in the study were
selected from the over 17,000 notifications and on-scene responses to
crashes in NASS sample jurisdictions.
[F] MVOSS uses two questionnaires that cover different subject areas,
with a small number of shared questions. Each questionnaire is
administered to a national sample of approximately 6,000 respondents,
the samples are independently drawn, and the same methods are used to
draw the samples. Essentially, MVOSS is two surveys.
[G] The vehicle identification number (VIN) indicates new equipment if
standard on the make and model; the dataset does not include the full
VIN, which is needed to check manufacturers or dealers for optional
equipment on the vehicle.
[H] NASS CDS and NMVCSS collect the vehicle identification number from
the vehicle inspection or the police accident report.
[I] Field inspection of crash scene, vehicles involved, interviews with
involved persons, and medical records for injured occupants.
[J] Field inspection of crash scene, vehicles involved, and interviews
with involved persons, focused on factors that led up to the crash.
[K] Event data recorder information is collected when available and
permission received.
[End of table]
[End of section]
Appendix VII: In-Vehicle Crash Avoidance Technologies and New NCAP:
NHTSA began the New Car Assessment Program (NCAP) in 1978 to provide
consumer information to the public. NCAP's goals are to improve
occupant safety by providing market incentives for vehicle
manufacturers to voluntarily design vehicles with improved
crashworthiness and provide independent information to aid consumers in
making comparative vehicle purchase decisions. To measure the relative
safety of new vehicles, NHTSA developed a series of crash tests that
would indicate a vehicle's relative crashworthiness.
For several years, NHTSA used only frontal crash tests and provided
data to consumers on the results. Subsequently, to provide information
to consumers that would be more easily understood, NHTSA developed the
five-star rating program. NHTSA used this rating method first in model
year 1994, and it is still in use today. The five-star program gives
new vehicles a rating on a one-to-five-star scale representing how well
a particular vehicle performed in the crash tests. Beginning in model
year 1997, NHTSA implemented side impact crash tests. In model year
2001, NHTSA incorporated rollover crashes into NCAP's testing and
ratings.
The Secretary of Transportation credits NCAP with encouraging major
safety improvements in the design of new vehicles. According to NHTSA,
recent advances in crash avoidance technology offer a new opportunity
for NCAP to inform consumers about new systems. In January 2007, NHTSA
issued an agency report entitled The New Car Assessment Program:
Suggested Approaches for Future Program Enhancements and requested
comments on options for enhancing NCAP, including information on crash
avoidance technologies. The proposal described various options for how
to present such information to consumers and indicated that information
differentiating the effectiveness of various new technologies might
have to wait for future evidence development.
In considering NHTSA's proposal for expanding NCAP, automobile
manufacturers and various stakeholders have emphasized the desirability
of (1) a rating system reflecting expected safety benefits, (2) efforts
to develop better evidence on new technologies, and (3) programs that
educate consumers about the nature of these technologies and how they
work.
After receiving and evaluating comments through a public hearing and
docket submissions, NHTSA issued its final notice in July 2008.
[Footnote 100] The notice summarized the comments received on the
agency report and provided the agency's decision on how it would
proceed with changes to NCAP. It was decided that for model year 2010,
the agency would implement a new crash avoidance program that would
communicate whether ESC, FCW, or LDW are standard or optional on
vehicles. NHTSA stated that such a rating system should be established
for two reasons: (1) to draw a greater distinction for consumers
regarding vehicles that are being equipped with ESC during the phase-in
period and (2) to provide an incentive for the accelerated deployment
of ESC and other new safety technologies that could help drivers
prevent severe and frequent crashes. According to NHTSA, ESC, FCW, and
LDW are the only technologies mature enough for inclusion in a crash
avoidance rating program at this time. All three address a major crash
problem, have had safety benefit projections, and have performance
tests and procedures available to ensure an acceptable performance
level.
NHTSA also notes that the agency will continue to seek public input on
the appropriateness of changes to the rating system or technologies,
and it anticipates using similar criteria for determining technologies
to include in the future.
It is not certain how further information on in-vehicle crash avoidance
technologies is to be obtained. NHTSA officials stated that they have
yet to determine the details of how they will test each new crash
avoidance technology and whether this will be done through an FOT or
other research.[Footnote 101] Similarly, with respect to future
generations of existing technologies (for example, improved versions of
curve speed warning), NHTSA has not indicated whether or how it might
obtain updated evidence on the quantifiable benefits of these
technologies. Finally, since many crash avoidance technologies may
interact with other safety features, it is not clear how a test would
resolve the overall safety impact of a particular technology, such as
LDW or curve speed warning.
In terms of consumer awareness of new technologies, the agency has
discussed focus group results in which NHTSA noted that "participants
may not fully grasp the importance of" new crash avoidance
technologies. NHTSA notes its agreement with points made by some
commenters in support of consumer education materials, a database of
nonagency sources of credible vehicle safety information, and
suggestions that the agency provide additional information at the point
of sale. According to NHTSA, it "continuously investigates ways to
improve marketing the NCAP vehicle ratings program...and will place the
results of…enhanced marketing studies in Docket No. NHTSA-02004-19104,
as completed."
[End of section]
Appendix VIII: Unintended Consequences:
When multiple trends converge, the potential future consequences may
differ from what was intended or anticipated from considering a single
trend in isolation or within the current context (see figure 20). In
some cases, future interactions between trends could result in
unintended consequences, potentially diluting the impact of positive
trends or exacerbating threats.[Footnote 102]
Figure 20: Unintended Consequences of the Interaction of Multiple
Trends:
[Refer to PDF for image}
This figure contains the following information as well as two
illustrations of the information:
Trends:
Crash avoidance + older drivers;
Possible interactions: Older drivers may be helped by crash avoidance
technologies such as backup warning systems illustrated below.
Unintended consequence: Crash avoidance technologies, such as backup
warning systems and night vision assistance, may enhance mobility of
older drivers, who then travel more miles and, therefore, experience
more crashes and fatalities.
Trends:
Crash avoidance + cell phones;
Possible interactions: Crash avoidance technologies could mitigate
negative effects of drivers using cell phones or other distracting
devices.
Unintended consequence: Drivers using a portable touch-screen phone and
examining a dashboard screen image at the same time could be further
distracted. (Rear view on backup screen, as shown, is activated when
car is in reverse.)
Sources: AAA Foundation for Traffic Safety and GAO.
[End of figure]
[End of section]
Appendix IX: Literature Review's Suggestion: Evaluating State Laws:
The congressional conference report that led DOT to conduct a
literature review on distracted driving suggested that the results of
that review might help "focus the federal research effort."[Footnote
103] The resulting review indicated one direction for research that
focuses specifically on countermeasures: "evaluation of the
effectiveness of State distraction-related laws" (Ranney 2008, 22).
[Footnote 104] Elsewhere, the review stated that while most efforts to
control driver behaviors might have limited effectiveness for driver
lifestyle choices such as using a cell phone while driving, a key
exception might be state GDL programs banning cell phone use for new
drivers, which include teens 16 to 17 years old.[Footnote 105]
A recent evaluation of a single state's GDL cell ban (not conducted by
DOT) indicates that the ban did not change teen driver behavior in the
state studied (Foss and others 2008). The literature review's
suggestion that DOT evaluate state laws--and the possible focus on GDL
laws--is still relevant in that neither the literature review nor the
recent GDL evaluation discussed whether different states (1) may have
implemented GDL cell bans in different ways or (2) are considering new
approaches to GDL implementation in the near future.
* Thus far, according to NHTSA, 19 states and the District of Columbia
have implemented GDL cell bans. The literature review did not indicate
whether some states currently (1) enforce the cell phone ban by, for
example, focusing on areas near high schools; (2) aim publicity about
the law at teens and parents; or (3) sponsor related programs to
encourage teen compliance or parental oversight.[Footnote 106]
(According to the report on the one state that was evaluated, that
state did not emphasize such approaches.)
* As we have documented throughout this report, fast-changing new
technologies are affecting highway safety. One objective of further
research could be to describe whether states with GDL cell bans are
using, encouraging, or considering for new technologies the future,
such as devices that could help police detect ongoing calls in passing
cars or in-car equipment to track, record, and report--to either
parents or police--driver use of portable phone use.[Footnote 107]
* A researcher in this area told us that no "poll" of states has
described how they are implementing GDL programs.
Thus, it is possible that some states are implementing or considering
the implementation of GDL cell bans differently from the state that was
tested. Research to evaluate the different ways that such bans are
implemented would be relevant to the suggestion in the literature
review.
[End of section]
Appendix X: Comments from the Department of Transportation:
Department of Transportation:
Office of the Secretary of Transportation:
Assistant Secretary for Administration:
1200 New Jersey Avenue, SE:
Washington, DC 20590:
September 26, 2008:
Ms. Katherine Siggerud:
Managing Director, Physical Infrastructure:
Ms. Nancy Kingsbury:
Managing Director, Applied Research and Methods:
U.S. Government Accountability Office:
441 G Street, NW:
Washington, DC 20548:
Dear Ms. Siggerud and Ms. Kingsbury:
The Government Accountability Office's (GAO) draft report on foresight
issues facing the Department of Transportation presents an interesting
theoretical approach, which attempts to extrapolate trends and identify
early intervention points to improve long term outcomes 20 or more
years hence. While these methods have laudable intent, they suffer from
several shortcomings. Implementing these approaches to long term issues
would require the diversion of substantial resources from ongoing
efforts with proven outcomes in an attempt to anticipate and define
future trends that may, or may not, develop. The National Highway
Traffic Safety Administration (NHTSA), for example, has honed a data-
driven, performance-oriented approach to addressing the issues most in
need of attention now and for the foreseeable future. These include
increasing the use of passenger restraint systems, reducing impaired
driving through behavioral and technological improvements, reducing
motorcycle fatalities, enhancing vehicle crash avoidance and
crashworthiness systems, and increasing consumer information to provide
for knowledgeable choices when purchasing a vehicle. These are the
primary actions that reflect NHTSA's focus on those areas that pose the
greatest risk on the Nation's highways and require continued and
increasingly effective action. While the Department is not unconcerned
about the issues that may emerge over the next 20 years, it must, in a
resource-constrained environment, focus its attention on evidence-based
strategies to improve safety.
The anticipatory mechanisms discussed in the draft report suffer from
methodological problems as well. First and foremost is their inability
to effectively contend with discontinuities that inevitably occur
within the extended timeframe during which they are intended to
operate. It is indeed the proverbial bus that you do not see coming
that poses the greatest danger of collision. There will certainly be
numerous discontinuities over the next 20 years, e.g., world events and
unanticipated technologies that will pose substantial and unpredictable
influence over what will become future key transportation safety
issues. For example the oil shocks of the 1970's were unanticipated and
resulted in substantial changes to the Nation's rolling stock as well
as new safety issues which could not have been anticipated 20 years
prior. In addition to these shortcomings, these techniques also require
massive amounts of data, sophisticated technical analytical tools,
skilled practitioners, and a significant investment of resources.
Further, these tools do not at this time appear to produce results with
the rigor necessary to form the basis of regulatory action by the
Department. These methods may well merit additional attention and
development in academia, which may be most adept at developing the
methods to their full potential.
The Department is looking towards the future to anticipate and address
major safety issues. The Federal Highway Administration and NHTSA's
actions to address the special needs of older drivers are a case in
point. While the methodology may differ from that offered in the GAO
draft report, the Department has initiated, together with its state
partners, actions intended to address the increasing population of
older drivers. The Department's efforts to meet this challenge were
recognized in GAO's recent report on the topic.[Footnote: see below]
That report describes both Federal efforts to implement safety measures
to address the special needs of older drivers and the difficulty states
face in committing scarce resources in anticipation of, rather than in
response to, significant safety concerns.
Finally, the GAO report is inaccurate in its conclusion that DOT has
not initiated research or action to counter the threat of electronic
driver distractions. NHTSA continues to amass a body of knowledge
regarding driver distractions, most recently issuing "Driver
Distraction: A Review of the Current State-of-Knowledge" in April, and
"Driver Strategies for Engaging in Distracting Tasks Using In-Vehicle
Technologies" in March. These are just the latest of many research
studies listed on NHTSA's internet website that are focused on the
topic of driver distraction. These are all part of NHTSA's efforts to
determine the nature and extent of the safety problem as well as
intervention strategies.
We appreciate the opportunity to offer comments on the draft report. In
addition to this statement for inclusion in the report, we have
separately provided a number of technical comments for GAO's
consideration in finalizing the document. Please contact Martin Gertel,
Director of Audit Relations, on 202-366-5145 with any questions.
Sincerely:
Signed by:
Linda J. Washington:
[Footnote: Older Driver Safety: Knowledge Sharing Should Help States
prepare for Increase in Older Driver Population," GAO-07-413, April
2007.]
[End of section]
Appendix XI: GAO Contacts and Staff Acknowledgments:
GAO Contacts:
Nancy R. Kingsbury, Managing Director, Applied Research and Methods,
(202) 512-2700 or kingsburyn@gao.gov. Katherine A. Siggerud, Managing
Director, Physical Infrastructure Issues, (202) 512-2834 or
siggerudk@gao.gov.
Staff Acknowledgments:
In addition to the individuals named above, Judith A. Droitcour,
Assistant Director; Nancy J. Donovan, Eric M. Larson, Joshua Ormond,
Penny Pickett, Beverly Ross, Raymond Sendejas, and Tina Won Sherman
made significant contributions to this report.[Footnote 108]
Bert Japikse and Andrea J. Levine, Office of General Counsel, provided
legal assistance.
[End of section]
References:
Abdel-Aty, Mohamed. 2003. Investigating the Relationship between
Cellular Phone Use and Traffic Safety. Institute of Transportation
Engineers. ITE Journal 73, no. 10 (October):38-42.
Asmeron, Ariam, and TaMara McCrae. 2006. The Evolution of Advanced
Research. Public Roads 69:6 (May-June):8-13.
Baker, Susan P., Li-Hui Chen, and Guohua Li. 2007. Nationwide Review of
Graduated Drivers Licensing. Washington, D.C.: AAA Foundation for
Traffic Safety, February.
Band, David, and Mike Perel. 2007. Senior Mobility Series. Article 8.
New Vehicle Technologies May Help Older Drivers. Public Roads 71, no. 3
(November/December):10-19.
Blumberg, Stephen J., and Julian V. Luke. 2008. Wireless Substitution:
Early Release of Estimates from the National Health Interview Survey,
July-December 2007. Atlanta: National Center for Health Statistics,
Centers for Disease Control and Prevention, May 13.
Brecher, Aviva, 2007. A Safety Roadmap for Future Plastics and
Composites Intensive Vehicles. Washington, D.C.: Department of
Transportation. DOT HS 810 863.
Brennan, Joe, Eyal Adi, and Bill Campbell. 2008. Trinity-Noble's
Celltinel® Solution: Guardian Angel®. PowerPoint presentation (July 23)
at [hyperlink, http://www.trinitynoble.com].
Burgett, August, Gowrishanker Srinivasan, and Raja Rangunathan. 2008. A
Methodology for Estimating Potential Safety Benefits for Pre-Production
Driver Assistance Systems. Washington, D.C.: NHTSA, May. Report DOT HS
810 945.
Chisholm, S. L., J. K. Caird, and J. Lockhart. Forthcoming. The Effects
of Practice with MP3 Players on Driving Performance. Accident Analysis
and Prevention.
Coates, Joseph. 1985. Foresight in Federal Government Policymaking.
Futures Research Quarterly (Summer):29-53.
Committee on National Statistics. 2005. Improving Data to Analyze Food
and Nutrition Policies. Washington, D.C.: National Academies Press.
Cooper, Peter J., and Yvonne Zheng. 2002. Turning Gap Acceptance
Decision-Making: The Impact of Driver Distraction. Journal of Safety
Research 33, no. 3 (Fall): 321-35.
Courtney, Hugh G., Jane Kirkland, and S. Patrick Viguerie. 1997.
Strategy Under Uncertainty. Harvard Business Review (November-
December): 67-79.
Dang, Jennifer N. 2007. Statistical Analysis of the Effectiveness of
Electronic Stability Control (ESC) Systems--Final Report. Washington,
D.C.: NHTSA, July. NHTSA Technical Report DOT HS 810 794.
Dewar, James. 2006. Improving Legislation on Long-Range Issues. NYU
Wagner School of Public Service, James Brademas Center for the Study of
Congress, Research Brief no. 4. New York, N.Y., December.
Dingus, Thomas A. 2007. The Impact of Driver Performance and Behavior
on Vehicular Safety and Crash Risk. Summary of presentation at
Preventing Crashes: Driver Safety through Human Factors Science,
seminar hosted by Federation of Behavioral, Psychological, and
Cognitive Sciences. U.S. Congress, U.S. House of Representatives,
Washington, D.C., November 13.
DOT (U.S. Department of Transportation). 1997. NHTSA 2020 Report:
People Saving People on the Road to a Healthier Future. Washington,
D.C.: September.
DOT (U.S. Department of Transportation). 2000a. The Changing Face of
Transportation. Washington, D.C.: Bureau of Transportation Statistics.
BTS00-007.
DOT (U.S. Department of Transportation). 2000b. Transportation Decision
Making: Policy Architecture for the 21st Century. Washington, D.C.
DOT (U.S. Department of Transportation). 2005. Research, Development,
Demonstration, and Deployment Roadmap for Hydrogen Vehicles and
Infrastructure to Support a Transition to a Hydrogen Economy.
Washington, D.C.: October.
DOT (U.S. Department of Transportation). 2007. U.S. Department of
Transportation Action Plan to Reduce Motorcycle Fatalities. Washington,
D.C.: October. DOT HS 810 855.
DOT (U.S. Department of Transportation). 2008a. DOT 1.0 Fatality Rate
Goal. Washington, D.C.: March.
DOT (U.S. Department of Transportation). 2008b. Refocus. Reform. Renew:
A New Transportation Approach In America. Washington, D.C.: July 29.
DOT (U.S. Department of Transportation). 2008c. Transportation Vision
for 2030. Washington, D.C.: Research and Innovative Technology
Administration, January.
FHWA (Federal Highway Administration). 2003. Managing and Organizing
Comprehensive Highway Safety in Europe. Washington, D.C.: April. FHWA-
PL-03-006:
FHWA (Federal Highway Administration). 2005. Roadway Human Factors and
Behavioral Safety. Washington, D.C.: May. FHWA-PL-05-005.
FHWA (Federal Highway Administration). 2006a. Halving Roadway
Fatalities: A Case Study from Victoria, Australia 1989-2004.
Washington, D.C.: April. FHWA-PL-06-011:
FHWA (Federal Highway Administration). 2006b. International Technology
Scanning Program: Safety Applications of Intelligent Transportation
Systems in Europe and Japan. Washington, D.C.: January. FHWA-PL-06-001.
FHWA (Federal Highway Administration). 2006c. Traffic Safety
Information Systems International Scan: Strategy Implementation White
Paper. McLean, Va.: September. FHWA-HRT-06-099.
Fine, Charles H. 1998. Clockspeed: Winning Industry Control in the Age
of Temporary Advantage Cambridge, Mass.: Perseus Books Group.
Fisher, Erik, Roop L. Mahajan, and Carl Mitcham. 2006. Midstream
Modulation of Technology: Governance from Within. Bulletin of Science,
Technology and Society 26, no. 6 (December): 485-96.
Foss, Robert D., and others. 2008. Short-Term Effects of a Teenage
Driver Cell Phone Restriction. Arlington, Va.: Insurance Institute for
Highway Safety, May.
FURORE. 2003. R&D Technology Roadmap: A Contribution to the
Identification of Key Technologies for a Sustainable Development of
European Road Transport. Graz, Austria: European Automotive Research
Partners Association. Available at [hyperlink, http://www.furore-
network.com].
GAO. 1997. Agencies' Strategic Plans under GPRA: Key Questions to
Facilitate Congressional Review. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO/GGD-10.1.16]. Washington, D.C.: May 1.
GAO. 2005a. 21st Century Challenges: Reexamining the Base of the
Federal Government. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-
05-325SP]. Washington, D.C.: February 1.
GAO. 2005b. Vehicle Safety: Opportunities Exist to Enhance NHTSA's New
Car Assessment Program. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-05-370]. Washington, D.C.: April 29.
GAO. 2007. A Call for Stewardship: Enhancing the Federal Government's
Ability to Address Key Fiscal and Other 21st Century Challenges.
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-93SP]. Washington,
D.C.: December 17.
GAO. 2008a. Highlights of a Forum Convened by the Comptroller General
of the United States: Strengthening the Use of Risk Management
Principles in Homeland Security. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-08-627SP]. Washington, D.C.: April 15.
GAO. 2008b. Traffic Safety: Grants Generally Address Key Safety Issues,
Despite State Eligibility and Management Difficulties. [hyperlink,
http://www.gao.gov/cgi-bin/getrpt?GAO-08-398]. Washington, D.C.: March
14.
GAO. 2008c. Traffic Safety: Improved Reporting and Performance Measure
Would Enhance Evaluation of High-Visibility Campaigns. [hyperlink,
http://www.gao.gov/cgi-bin/getrpt?GAO-08-477]. Washington, D.C.: April
25.
GAO. 2008d. Traffic Safety: NHTSA's Improved Oversight Could Identify
Opportunities to Strengthen Management and Safety in Some States.
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-788]. Washington,
D.C.: July 14.
GHSA (Governors Highway Safety Association). 2003. GHSA Joins AAA in
Announcing Distracted Driving Research. Press release, National Press
Club, Washington, D.C., August 6.
Ginsburg, Kenneth, and others. 2008. National Young Driver Survey: Teen
Perspective and Experience with Factors That Affect Driving Safety.
Pediatrics 121, no. 5 (May): e1391-e1403.
Glassbrenner, Donna, and Tony Jianqiang Ye. 2007. Driver Cell Phone Use
in 2006--Overall Results. Traffic Safety Facts: Research Note.
Washington, D.C.: NHTSA, July. DOT HS 810 790.
Grim, Terry, and Scott Reif. 2008. Assessing Foresight Capabilities: An
Organizational Scorecard. Presented to World Future Society, Bethesda,
Maryland, July 27.
Hahn, Robert W., and James E. Prieger. 2006. The Impact of Driver Cell
Phone Use on Accidents. B. E. Journals in Economic Analysis and Policy:
Advances in Economic Analysis and Policy 6, no. 1:1-37.
Harvard Business School. 2005. Strategy: Create and Implement the Best
Strategy for Your Business. Cambridge, Mass.: Harvard Business
Publishing.
Horrey, William J., and Christopher D. Wickens. 2006. Examining the
Impact of Cell Phone Conversations on Driving Using Meta-Analytic
Techniques. Human Factors 48, no.1 (Spring):196-205.
Hosking, Simon, and others. 2006. The Effects of Text Messaging on
Young Novice Driver Performance. Victoria, Australia: Monash University
Accident Research Centre, February.
Hunton, James, and Jacob M. Rose. 2005. Cellular Telephones and Driving
Performance: The Effects of Attentional Demands on Motor Vehicle Crash
Risk. Risk Analysis (Oxford) 25, no. 4 (August): 855-66.
INTOSAI (International Organization of Supreme Audit Institutions).
2007. Program Evaluation for SAIs: A Primer. Draft. Washington, D.C.:
November.
Kahane, Charles J. 2004. Lives Saved by the Federal Motor Vehicle
Safety Standards and Other Vehicle Safety Technologies 1960-2002:
Passenger Cars and Light Trucks with a Review of 19 FMVSS and Their
Effectiveness in Reducing Fatalities, Injuries, and Crashes.
Washington, D.C.: NHTSA, October. Report DOT HS 809 833.
Kanianthra, Joseph. 2007. United States Government Status Report.
Proceedings of the 20th International Technical Conference on the
Enhanced Safety of Vehicles. France: Lyon.
Klauer, S. G., and others. 2006. The Impact of Driver Inattention on
Near-Crash/Crash Risk: An Analysis Using the 100-Car Naturalistic
Driving Study Data. Washington, D.C.: U.S. Department of
Transportation, April. DOT HS 810 594.
Lange, Robert C. 2007. Executive Director, Vehicle Structure and Safety
Integration, General Motors North America, Warren, Michigan. Letter to
Nicole R. Nason, Administrator, National Highway Traffic Safety
Administration, Washington, D.C., April 25.
Lerner, Neil, and Gina Balliro. 2003. Driver Strategies for Engaging in
Distracting Tasks Using In-Vehicle Technologies: Focus Group Report.
Washington, D.C.: NHTSA. Interim report under contract dtnh22-99-d-
07005.
Lerner, Neil, and Shelly Boyd. 2004. On-Road Study of Willingness to
Engage in Distracting Tasks. Washington, D.C.: NHTSA.
Lerner, Neil, Jeremiah Singer, and Richard Huey. 2008. Driver
Strategies for Engaging in Distracting Tasks Using In-Vehicle
Technologies. Washington, D.C.: NHTSA, March. DOT HS 810 919.
McCartt, Anne T., Laurie Helinga, and Kel Bratiman. 2006. Cell Phones
and Driving: Review of Research. Traffic Injury Prevention 7:89-106.
McEvoy, Suzanne, and others. 2005. Role of Mobile Phones in Motor
Vehicle Crashes Resulting in Hospital Attendance: A Case-Crossover
Study. British Medical Journal (International Edition: London) 331, no.
7514 (August 20-27):428-30.
NHTSA (National Highway Traffic Safety Administration). 2008a. Driver
Electronic Device Use in 2007. Traffic Safety Facts Research Note.
Washington, D.C.: June. DOT HS 810 963.
NHTSA (National Highway Traffic Safety Administration). 2008b.
Countermeasures That Work: A Highway Safety Countermeasure Guide for
State Highway Safety Offices, 3rd ed. Washington, D.C.: January. DOT HS
810 891.
NHTSA (National Highway Traffic Safety Administration). 2008c.
Statistical Analysis of Alcohol-Related Driving Trends: 1982-2005.
Washington, D.C.: May. DOT HS 810 942.
Newman, Betty. 2008. UConnect Web Brings Internet on the Road. June 26.
Available at [hyperlink,
http://blog.chryslerllc.com/blog.do?p=entry&id=403].
Paulo, Loureiro, Adolfo Sachsida, and Tito Moreira. 2004. Traffic
Accidents: An Econometric Investigation. Economics Bulletin 18, no. 3:
1-7.
Popper, Steven W., Robert J. Lempert, and Steven C. Bankes. 2005.
Shaping the Future: Scientific Uncertainty Often Becomes an Excuse to
Ignore Long-term Problems Such As Climate Change. It Doesn't Have to Be
So. Scientific American 292, no. 4 (April):66-71.
Rakauskas, Michael E., Leo J. Gugerty, and Nicholas J. Ward. 2004.
Effects of Naturalistic Cell Phone Conversations on Driving
Performance. Journal of Safety Research 35, no. 4:453-64.
Ranney, Thomas A. 2008. Driver Distraction: A Review of the Current
State of Knowledge. Prepared for DOT. Washington, D.C: April. NTIS
report DOT HS 810 704.
Redelmeier, Donald A., and Robert J. Tibshirani. 1997. Association
between Cellular-Telephone Calls and Motor Vehicle Collisions. The New
England Journal of Medicine 336, no. 7 (February 13): 453-58.
Rejeski, David. 2003. S&T Challenges in the 21st Century: Strategy and
Tempo. Pp. 47-57 in AAAS Science and Technology Policy Yearbook 2003,
edited by Albert H. Teich and others. Washington, D.C.: American
Association for the Advancement of Science.
Rejeski, David, and Carly Wobig. 2002. Long-term Goals for Governments.
Foresight 4, no. 6:14-22.
Rosenker, Mark V. 2007. Chairman, NTSB, Opening Remarks at the
Northwestern University Patterson Transportation Lecture Series,
Evanston, Illinois, April 11.
Schattler, Kerrie L., and others. 2006. Assessing Driver Distractions
from Cell Phone Use While Driving: A Simulator-Based Study. Paper
submitted at the 85th Annual Meeting of the Transportation Research
Board, Washington, D.C.: January.
Seo, Dong-Chul, and Mohammad R. Torabi. 2004. The Impact of In-Vehicle
Cell-Phone Use on Accidents or Near-Accidents among College Students.
Journal of American College Health 53, no.3 (November/December): 101-
07.
Stanton, Michael. 2007. Chief Executive Officer, Association of
International Automobile Manufacturers, personal communication, June
12.
Strayer, David L., and Frank Drews. 2004. Profiles in Driver
Distraction: Effects of Cell Phone Conversations on Younger and Older
Drivers. Human Factors 46, no. 4 (Winter):640-49.
Strayer, David L., Frank A. Drews, and Dennis Crouch. 2006. A
Comparison of the Cell Phone Driver and the Drunk Driver. Human Factors
48, no. 2 (Summer):381-91.
Svenson, Ola, and Christopher J. D. Patten. 2005. Mobile Phones and
Driving: A Review of Contemporary Research. Cognitive Technology
Journal 7:182-97.
Vlingo Corporation. 2008. Consumer Text Messaging Habits. Cambridge,
Mass.: May.
VTRC (Virginia Transportation Research Council). 2005. VTRC a Co-
sponsor of Groundbreaking Study of Driver Behavior. Press release,
August 15. Accessible at [hyperlink,
http://vtrc.virginiadot.org/BriefDetails.aspx?Id=19].
Walker, David M., Comptroller General. 2005. Social Security Reform:
Early Action Would Be Prudent. GAO-05-397T. Washington, D.C.: March 9.
Walker, David M., Comptroller General. 2007. Enhancing Performance,
Accountability, and Foresight. GAO-07-165CG. Speech before the
Caribbean Organization of Supreme Audit Institutions (CAROSAI), St.
Lucia, West Indies.
Wochinger, Kathryn, and others. 2008. Human Factors Forum on Advanced
Vehicle Safety Technologies: Summary and Proceedings. Washington, D.C.:
DOT/National Highway Traffic Safety Administration, March. DOT HS 810
918.
Younes, Maged. 2005. Action Implementation and Evaluation. Presentation
at the World Health Organization Workshop on Guiding Public Health
Policy in Areas of Scientific Uncertainty, University of Ottawa,
Ottawa, Canada, July 11-13.
[End of section]
Footnotes:
[1] Exercising foresight consists of basing policies on an
understanding of forces shaping the future (see Coates 1985, 33; full
citations are provided in the reference list at the end of this
document). In this context, a potentially significant trend is one
that, although somewhat uncertain, may substantially affect progress
toward basic goals across a time horizon more than 5 years forward. (As
a point of comparison, the Government Performance and Results Act of
1993 (GPRA) requires that agencies' strategic plans cover a period not
less than 5 fiscal years forward.)
[2] We developed this definition from the concept of "clockspeed" as a
characteristic of product-development, introduced by Fine (1998).
[3] Michael Stanton (Stanton 2007), Chief Executive Officer,
Association of International Automobile Manufacturers (AIAM), expressed
a similar view with respect to automotive developments.
[4] Risk management and strategic planning literature emphasizes the
need for organizations to recognize and address both (1) opportunities-
-that is, potential solutions (or partial solutions) to one or more
problems and (2) threats, each of which represents a potential problem.
Within the context of this report, a trend presents an opportunity for
improving future safety if it promises to reduce future fatalities (or
to further reduce them) if DOT successfully promotes, encourages, or
helps develop that trend.
[5] Each technology-based trend in this list combines (1) an evolving
series of technology developments and new products with (2) consumer
adoption and patterns of use.
[6] In contrast, GPRA requires that agencies' strategic plans cover a
period not less than 5 fiscal years forward.
[7] In some cases, it may be appropriate to consider a technology-based
trend as a whole (for example, electronic driver distractions); in
other cases, it may be more appropriate to focus separately on various
types of products or the specific impacts associated with them (for
example, differentiating driver voice calls from driver text
messaging).
[8] Trajectory refers to the path of a trend's forward movement and
future phases (see Fisher, Mahajan, and Mitcham 2002). A trend's impact
can change over time, either increasing or decreasing. Such changes can
occur at various rates--for example, increasing at an exponential rate
or decreasing at a slow, incremental rate.
[9] For in-vehicle crash avoidance technologies, we focus specifically
on those introduced after electronic stability control (ESC).
Introduced in the mid-1990s, ESC improves vehicle stability when a
driver is in danger of crashing; for example, ESC helps prevent loss of
control--which can lead to run-off-the-road rollovers--by applying
brakes differentially to each wheel to increase stability. DOT has
required ESC in all passenger vehicles manufactured after September 1,
2011. (Our focus on crash avoidance technologies does not address
commercial vehicles or roadway systems that do not involve in-vehicle
technology.)
[10] We use highway safety to refer to safety on any public road.
[11] According to DOT, VMT measurement should be improved; for example,
current procedures may underestimate motorcycle travel and do not
measure pedestrian exposure.
[12] The rate most recently reported, in August 2008, was 1.38
fatalities per 100 million VMT for 2007.
[13] NHTSA recently told us that it should meet the 1.0 fatality rate
for occupants of passenger cars and that the higher overall rates
reflect fatalities involving pedestrians, motorcyclists, and heavy
trucks, as well as those for passenger-car occupants. NHTSA also said
that more accurate measures of exposure are needed.
[14] The fatality rate can be calculated as the number of fatalities
divided by VMT. If such rates decline, the number of fatalities may or
may not decline, depending on trends in VMT.
[15] VMT is currently declining as fuel costs rise, but whether the
decline will continue through future years is uncertain.
[16] Roughly half of all fatalities occur for potentially vulnerable
road-user groups, defined as including elderly persons, pedestrians,
bicyclists, motorcyclists, occupants of compact cars (who may be
vulnerable in crashes involving larger vehicles), and occupants of
other cars when encountering heavy trucks. (We discuss trends in
vulnerable groups in appendix IV.)
[17] A different kind of new safety technology that carmakers are
including in some vehicles is automatic crash notification. Private
sector companies are also developing other new technologies for
surveillance or enforcing safety rules. These include systems to more
effectively enforce (1) speed limits and (2) laws on elevated blood
alcohol content (BAC).
[18] While automobile manufacturers introduce new model cars every 4 to
6 years, they introduce new options more frequently. Additionally,
major manufacturers are adjusting their strategies to better compete in
a "higher clockspeed" world (Fine 1998, 64 and 238-239).
[19] The length of time to reach a given percentage of cars on the road
depends on sales volumes for new vehicles with the technology in each
model year. If manufacturers install a particular new technology in
only luxury models at first, it takes longer for it to reach a majority
of vehicles on the road.
[20] A limited toll collection application (EZ pass) operational in
some locations allows drivers with prepaid accounts to pass through
toll booths without stopping.
[21] Also referred to as a haptic warning, one example of a tactile
warning is vibration on one side of the driver's seat.
[22] Backup cameras positioned in the rear of the car transfer images
to a dashboard screen when the car is in reverse (two such screens are
illustrated in appendix V).
[23] Changes toward riskier driver behavior in response to new
technologies designed to increase safety have been variously termed
risk compensation, risk homeostasis, moral hazard, and the usability
paradox.
[24] Our recent reviews of NHTSA safety grants to states include GAO
2008b and 2008d.
[25] Information on selected crash datasets and driver-behavior
datasets maintained by NCSA is in appendix VI.
[26] TRB conducts the second Strategic Highway Research Program (SHRP
2), authorized by the Congress in 2005 to study comprehensive crash
causation, congestion, and many other topics. RITA is the liaison to
TRB for SHRP 2.
[27] For example, a particular trend's impact may be increasing at a
fast pace or changing very slowly. We define quantitative data as
"strong" depending on validity, reliability, and generalizability.
Validity refers to unbiased counts or measurement of what one intends
to measure; reliability concerns reproducibility and consistency (that
is, the achievement of similar results if a study were repeated using
the same procedures); generalizability refers to the applicability of
study results to the population of interest. An additional indication
of strength would be multiple studies or different sources of data.
[28] Organizational networking involves establishing relationships with
others within one's organization and in other organizations (in both
the public and private sectors) in order to exchange information and,
in some cases, influence debates or actions.
[29] DOT actions are limited by its authority and budget.
[30] An FOT is a test conducted under typical operating conditions for
the technology being tested. Each FOT discussed here involved
outfitting a limited number of cars with sensors and other equipment
and had a relatively small sample; for example, one used 78 drivers and
11 cars, with each driver using a vehicle for 4 weeks.
[31] NHTSA's estimates indicate that LDW and FCW technologies are,
respectively, 6-11 percent and 15 percent safety effective. In
comparison, NHTSA characterizes ESC as 59 percent safety effective.
[32] NHTSA will be using three criteria for including crash avoidance
technologies in NCAP--(1) the safety technology addresses a major crash
problem, (2) estimates or projections of safety benefits are available,
and (3) performance tests and procedures are in place to ensure
adequate performance for each labeled model. For the future, automobile
manufacturers have pointed to the importance of efforts to develop
better evidence on new technologies. (We discuss post-ESC in-vehicle
crash avoidance technologies in appendix VII.)
[33] We discuss issues related to crash avoidance and NCAP in appendix
VII.
[34] FMCSA and FHWA have established a parallel VII initiative called
Smart Roadside, to address the development of VII for the commercial
vehicle industry.
[35] RITA's Volpe Center is preparing a benefit-cost analysis on VII
deployment scenarios (scheduled for delivery in late 2008).
[36] RITA officials told us that the identities of individual vehicles
would not be recorded at any point in the data collection process.
[37] The VII system would be more complex than figure 12 shows, because
it involves GPS satellite communications to identify a vehicle's
location and possibly other technologies, such as wireless Internet or
other modes of communication, as well as technology to allow
interoperability across subsystems.
[38] The goal of CICAS is to reduce accidents by alerting drivers when
they or other vehicles are projected to violate traffic control devices
and advising drivers about "gap acceptance" when deciding to maneuver
through an intersection (after making a legal stop) or to make left
turns. EVSCA is a NHTSA initiative to evaluate whether the
effectiveness of vehicle-to-vehicle communications (either alone or in
combination with stand-alone crash-avoidance technologies) could
benefit from technologies allowing the vehicle to communicate with
roadside or other sensors.
[39] Various functions could be added to VII incrementally as new
technologies become available.
[40] In the 100-car study, video cameras and sensors were installed in
approximately 100 vehicles, each tracked for about a year of "normal
driving" within an 18-month period in 2003-2004. Designs for each of
the 13 primary studies are characterized by strengths and weaknesses
(see appendix III).
[41] The surveys were a Harris Interactive® poll and a Nationwide
Mutual Insurance survey.
[42] The total amount of driver voice phoning is higher because the 6
percent figure does not include driver use of hands-free equipment
(such as earpieces and speakerphones). The observations were obtained
in the National Occupant Protection Use Survey (NOPUS) (see appendix
VI).
[43] This questionnaire survey was based on landline telephone sampling
and interviewing.
[44] Hosking and others (2006) indicate that driver texting increases
glances away from the road to 40 percent of the time from a baseline of
10 percent.
[45] The self-report survey included responses from 4,800 cell phone
users.
[46] To further illustrate the cohort effect, by 2020 drivers now 18 to
28 years old will be 30 to 40 years old; those now 58 to 68 years old
will be 70 to 80.
[47] Motorcycle mounts and car sun visors with built-in DVD players are
also available.
[48] Safe, Accountable, Flexible, Efficient Transportation Equity Act:
A Legacy for Users, Pub. L. No. 109-59, § 2003(d) (2005).
[49] H.R. Rep. No. 109-307, at 181-182 (2005).
[50] 23 U.S.C. § 402(a).
[51] The GHSA campaign was to have been joined by research sponsored by
the American Automobile Association that analyzed potentially
distracting activities of 70 volunteer drivers for 3 hours per driver,
based on cameras placed in their vehicles. A range of distractions were
observed.
[52] GDL programs typically involve three stages--driving with
supervision, restricted driving without supervision (such as limited
night-driving or number of teen passengers), and unrestricted driving.
Minimum ages or other requirements are set for passing from one stage
to another. A NHTSA guide describes such programs (NHTSA, 2008b, 4-4
and 4-9).
[53] NHTSA also told us that data from ongoing or planned studies, such
as a large study using the basic methodology of the 100-car study,
might suggest directions for developing countermeasures. (We describe
the larger study--known as the 2,500-car study--later in this report.)
[54] CEA's driver education materials are displayed on its consumer Web
site at [hyperlink, http://www.DigitalTips.org].
[55] This statement is based on national polls conducted in 2007 and
2008 by Nationwide Mutual Insurance Company, Zogby International, and
Harris Interactive®, as well as other regional and state polls
conducted in the same time period.
[56] One possible source of information on new and future directions in
technology development would be increased networking with associations
or other sources in the electronics industry.
[57] A risk management approach can be broadly defined as "a strategic
process for helping policymakers make decisions about assessing risk,
allocating finite resources, and taking actions under conditions of
uncertainty" (GAO 2008a, 1).
[58] With respect to high-stakes situations of this type, the
precautionary principle, which is widely applied in the European Union,
recognizes that government intervention beyond that normally justified
by scientific evidence may be warranted if there are signals that a
possible threat may, if unchecked, seriously harm the population.
[59] At the same time, some risk management experts have noted that
such criteria may be appropriately balanced against the possibility
that a trend may not develop as anticipated. For example, a trend that
appears likely to carry a substantial future threat may, in time, be
mitigated by unforeseen developments.
[60] Strong quantitative evidence is defined by figure 9 as Level 3
evidence.
[61] Validity refers to unbiased counts or measurement of what one
intends to measure, reliability concerns reproducibility and
consistency (that is, the achievement of similar results if the study
were repeated using the same procedures), and generalizability refers
to the applicability of study results to the population of interest.
[62] Recent research, such as a study by the Committee on National
Statistics (2005), has recognized timeliness as a strength of high-
quality policy-relevant data.
[63] We do not consider evidence development on VII here because it
would be premature.
[64] The analysis separated crash data for vehicles with and without
ESC and then, for these two vehicle groups, compared ratios of fatal
crashes of the type addressed by ESC to other fatal crashes.
[65] We describe these and other datasets maintained by NCSA in
appendix VI.
[66] We outline selected NCSA crash datasets in appendix VI.
[67] The NOPUS dataset is described in appendix VI.
[68] The percentage of adults living in cell-phone-only households has
been steadily increasing (Blumberg and Luke 2008).
[69] With small samples, it is difficult to determine whether patterns
observed occurred by chance alone.
[70] One caveat is that the results of this study may not be
generalizable to the entire United States, in part because the study
will be conducted in three or four geographic areas.
[71] For selected analyses, analysts will compare two sets of results
from the 2,500-car study: those based on actual crashes and those based
on near misses.
[72] Differences between the design of the 100-car pilot study and the
2,500-car study--mostly improvements instituted in the 2,500-car study-
-make it difficult to develop comparisons of data from the two. An
additional future study, designed to be comparable to the 2,500-car
study, would be needed to validly track change over time.
[73] A TRB official stated that this study, as originally proposed, was
to include 4,000 cars tracked over 3 years but, for budgetary reasons,
was cut back to 2,500 cars tracked for 2 years.
[74] For example, barrier crash tests are not suited to assessing crash
avoidance technologies, and the crash datasets are not suited to
achieving valid counts of crashes in which drivers were using
electronic devices, such as hands-free phones or texting.
[75] For example, NOPUS roadside observations of electronic driver
distractions worked for driver use of handheld portable phones but not
for more recent, more difficult to observe uses of electronic devices.
[76] We have not researched the privacy issues or solutions relevant to
such a tracking system.
[77] GPRA requires that strategic plans contain "an identification of
those key factors external to the agency and beyond its control that
could significantly affect the achievement of...goals and objectives."
Such factors, according to Office of Management and Budget Circular No.
A-11, may remain stable, change "within predicted rates," or vary to an
unexpected degree.
[78] Citing the tradeoff between safety and mobility, a DOT official
discussed the following examples: safety can benefit from congestion
because drivers are traveling at lower speeds; safety can suffer from
efforts to alleviate congestion by widening highways because such
modifications are conductive to higher speeds. (DOT also told us that
congestion can result in a high differential in speed, which could
cause crashes.)
[79] DOT planning and accountability materials include the strategic
plan, performance and accountability report, and budget proposal.
[80] This report, a collaborative effort by FHWA, FMCSA, NHTSA, and the
Office of the Secretary, discussed why DOT did not meet its 2008
fatality reduction goal and how its programs could achieve its 2011
goal (1 fatality per 100 million VMT). Four fatality submeasures were
established--passenger vehicle occupants, nonoccupants, motorcycle
riders, and large-truck and bus-related fatalities. The purpose of this
approach is to more closely examine the fatality rates of the different
segments of highway users and "devote greater energy and resources and
develop new strategies to combat sub-measure trends that are impeding
progress to the overall 1.0 goal." The report noted that "by isolating
fatalities by class of road user, the Department believes that it has
the greatest opportunity to develop appropriate new strategies to
address the factors and behaviors that cause each type of fatality."
[81] A 1997 effort specifically focused on the future of highway safety
that has not been updated is the NHTSA 2020 Report on trends that could
affect highway safety in the year 2020 (see DOT 1997).
[82] DOT also told us that RITA's Bureau of Transportation is
revisiting some of its long-term forecasts published in The Changing
Face of Transportation (DOT 2000a). By incorporating more recent annual
data, the bureau is reviewing the accuracy of the previous forecasts
and will update trends based on the new data.
[83] Internal report prepared for FMCSA.
[84] The report concluded that technological innovation will drive the
"next wave of change in the trucking industry." Such innovation
includes new vehicle technologies, new driver technologies, system
integration, expanded data exchange, and "more sophisticated"
statistical analysis of available data. In response, FMCSA should
determine whether it has staff possessing the technological knowledge
and skills to implement data collection, maintenance, and exchange
programs and to develop platforms to support such programs.
[85] Recent FHWA safety-related reports are FHWA 2003; 2005; 2006a;
2006b.
[86] This has ranged from attending technology-oriented meetings, such
as those hosted by the Society of Automotive Engineers and ITS America,
to the annual Lifesavers Conference, which reports on emerging
developments in behavior by members of high-risk groups. (ITS refers to
intelligent transportation systems.) A senior NHTSA official
responsible for research on behavioral issues described ongoing liaison
efforts with groups such as the Governors Highway Safety Association
and the Centers for Disease Control and Prevention. NHTSA officials
responsible for research on vehicle safety said they network
extensively with suppliers and automobile manufacturers, meeting
periodically at DOT headquarters to understand the potential of new
technology for vehicle safety.
[87] The four roads discussed are safety codes, standards, and
regulations; infrastructure development and deployment; safety
education, outreach, and training; and medium-and heavy-duty vehicle
development, demonstration, and deployment. The map for each road
includes four areas: anticipated long-term outcomes (11 to 20 years);
challenges and requirements; pathways, projects, and products; and
timelines.
[88] For example, the review of intelligent infrastructure systems
resulted in the Minister of State for Transport setting out next steps
for stakeholder testing of policies for robustness. The scenarios are
used to effectively manage long-term risks while taking advantage of
opportunities.
[89] For example, it is key to include an ongoing function and
involvement with foresight throughout an organization (Grim and Reif
2008).
[90] DOT told us that BTS's Trending and Forecast Team analyzes long-
term and short-term trends on key transportation indicators.
[91] Other ways of exercising foresight include, but are not limited
to, setting new goals for the future (such as the space program goal
set by President Kennedy), exploring potential consequences of a new
policy direction, and making quantitative projections of trends that
have been tracked over time.
[92] We also met with Delegate Jeff Waldstreicher, 18th Legislative
District, Montgomery County, Maryland, and we exchanged e-mails with
additional persons or groups, including Lisa Lewis of The Partnership
for Safe Driving, Washington, D.C.
[93] Association executives and staff also participated in these
meetings. At AIAM this included Michael J. Stanton, President and CEO,
and Michael X. Cammisa, Director of Safety. At the Auto Alliance, this
included Dave McCurdy, President and CEO, and Robert Strassburger, Vice
President, Vehicle Safety and Harmonization.
[94] Crash risk estimates are sometimes based on near misses or
simulated crashes. Only 1 of the 12 peer-reviewed primary studies
included texting and phoning and did not separate the two activities in
the analysis; one of the authors told us that texting was not as
prevalent as phoning. According to NHTSA, the 100-car study did not
include driver texting.
[95] This included one simulator study, two self-report surveys, and
two case-control record-check studies.
[96] The likelihood of a crash or near miss associated with a
particular driver behavior reflects a combination of (1) the frequency
with which drivers engage in that behavior and (2) the riskiness of the
behavior when engaged in.
[97] Additionally, a Brazilian study of crash risk compared smoking,
phoning, and the presence of children in the car.
[98] According to DOT, current measures of VMT underestimate motorcycle
use and better measures are needed.
[99] Although improved seat belts, air bags, and other safety features
help protect drivers and passengers in smaller cars, they are still
likely to be vulnerable in crashes.
[100] 73 Fed. Reg. 40,016 (July 11, 2008).
[101] The method NHTSA used to test ESC--comparisons of actual crashes
for cars with and without ESC, across multiple years of data--would be
likely to take 5 to 10 years from the time a new technology is
introduced in several new models. By then, automakers would be likely
to have introduced newer technologies.
[102] Another example is that a new development can be viewed as having
a positive effect, such as small cars lowering the cost of fuel for
consumers. However, their construction could compromise the safety
benefits of crashworthiness technologies. When considered in light of
the projected volume of commercial trucks on the nation's highways,
these interacting trends could lead to more fatalities in the future.
[103] H.R. Rep. No.109-307, at 181-182 (2005).
[104] The literature review also mentioned in-vehicle systems for
information and entertainment; it did not suggest evaluations of these
technologies as a possible research direction. Systems built into new
cars include OnStar™ and Sync™, among others. OnStar™ has reported that
driver calls made with its hands-free technology do "not increase the
risk of collision as compared to normal driving" (Lange 2007).
[105] Most states now have GDL programs, which have generally proven
effective (see NHTSA 2008b and Baker, Chen, and Li 2007).
[106] However, high-visibility enforcement has been used in other areas
of safety (see GAO 2008c, 24 and 26).
[107] Related technology is discussed in Brennan, Adi, and Campbell
(2008). DriveCam video-recording and other systems for oversight of
drivers can be used by parents of teens (see [hyperlink,
http://www.drivecam.com]).
[108] The following individuals contributed to this report during the
summer internship program with GAO: Grace E. Cho, Griffin Glatt-Dowd,
and Richard M. Todaro. Additionally, Charity J. Goodman, formerly with
GAO, contributed to an early phase of this study.
[End of section]
GAO's Mission:
The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people.
GAO examines the use of public funds; evaluates federal programs and
policies; and provides analyses, recommendations, and other assistance
to help Congress make informed oversight, policy, and funding
decisions. GAO's commitment to good government is reflected in its core
values of accountability, integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony:
The fastest and easiest way to obtain copies of GAO documents at no
cost is through GAO's Web site [hyperlink, http://www.gao.gov]. Each
weekday, GAO posts newly released reports, testimony, and
correspondence on its Web site. To have GAO e-mail you a list of newly
posted products every afternoon, go to [hyperlink, http://www.gao.gov]
and select "E-mail Updates."
Order by Mail or Phone:
The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or
more copies mailed to a single address are discounted 25 percent.
Orders should be sent to:
U.S. Government Accountability Office:
441 G Street NW, Room LM:
Washington, D.C. 20548:
To order by Phone:
Voice: (202) 512-6000:
TDD: (202) 512-2537:
Fax: (202) 512-6061:
To Report Fraud, Waste, and Abuse in Federal Programs:
Contact:
Web site: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]:
E-mail: fraudnet@gao.gov:
Automated answering system: (800) 424-5454 or (202) 512-7470:
Congressional Relations:
Ralph Dawn, Managing Director, dawnr@gao.gov:
(202) 512-4400:
U.S. Government Accountability Office:
441 G Street NW, Room 7125:
Washington, D.C. 20548:
Public Affairs:
Chuck Young, Managing Director, youngc1@gao.gov:
(202) 512-4800:
U.S. Government Accountability Office:
441 G Street NW, Room 7149:
Washington, D.C. 20548: