Crash Weighting Analysis
Crash Weighting Analysis
T.F. Scott Darling, III
Federal Motor Carrier Safety Administration
January 23, 2015
[Federal Register Volume 80, Number 15 (Friday, January 23, 2015)]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: 2015-01144]
DEPARTMENT OF TRANSPORTATION
Federal Motor Carrier Safety Administration
[Docket No. FMCSA-2014-0177]
Crash Weighting Analysis
AGENCY: Federal Motor Carrier Safety Administration (FMCSA), DOT.
ACTION: Notice; request for public comment.
SUMMARY: FMCSA announces a study to inform decision making about the
feasibility of using a motor carrier's role in crashes as an indicator
of future crash risk in response to stakeholder interest
and as part of the Agency's commitment to continuous improvement. This
study assesses (1) whether Police Accident Reports (PARs) provide
sufficient, consistent, and reliable information to support crash
weighting determinations; (2) whether a crash weighting determination
process would offer an even stronger predictor of crash risk than
overall crash involvement and how crash weighting would be implemented
in the Agency's Safety Measurement System (SMS); and (3) how FMCSA
might manage a process for making crash weighting determinations,
including the acceptance of public input. This notice advises the
public of the availability of the study report for review and comment,
along with a request for feedback on what steps the Agency should take
regarding crash and PAR data quality.
DATES: Comments must be received on or before February 23, 2015.
ADDRESSES: You may submit comments bearing the Federal Docket
Management System (FDMS) Docket ID FMCSA-2014-0177 using any of the
Federal eRulemaking Portal: Go to www.regulations.gov.
Follow the on-line instructions for submitting comments.
Mail: Docket Management Facility; U.S. Department of
Transportation, 1200 New Jersey Avenue SE., West Building Ground Floor,
Room W12-140, Washington, DC 20590-0001.
Hand Delivery or Courier: West Building Ground Floor, Room
W12-140, 1200 New Jersey Avenue SE., Washington, DC, between 9 a.m. and
5 p.m., ET, Monday through Friday, except Federal Holidays.
Each submission must include the Agency name and the docket number
for this notice. Note that DOT posts all comments received without
change to www.regulations.gov, including any personal information
included in a comment. Please see the Privacy Act heading below.
Docket: For access to the docket to read background documents or
comments, go to www.regulations.gov at any time or visit Room W12-140
on the ground level of the West Building, 1200 New Jersey Avenue SE.,
Washington, DC, between 9 a.m. and 5 p.m., ET, Monday through Friday,
except Federal holidays. The on-line Federal document management system
is available 24 hours each day, 365 days each year. If you want
acknowledgment that we received your comments, please include a self-
addressed, stamped envelope or postcard or print the acknowledgement
page that appears after submitting comments on-line.
Privacy Act: In accordance with 5 U.S.C. 553(c), DOT solicits
comments from the public to better inform its rulemaking process. DOT
posts these comments, without edit, including any personal information
the commenter provides, to www.regulations.gov, as described in the
system of records notice (DOT/ALL-14 FDMS), which can be reviewed at
FOR FURTHER INFORMATION CONTACT: For information concerning this study,
contact Ms. Dee Williams, Chief, Compliance Division, Federal Motor
Carrier Safety Administration, 1200 New Jersey Avenue SE., Washington,
DC 20590, Telephone 202-366-1812 or by email: email@example.com. If
you have questions on viewing or submitting material to the docket,
contact Docket Operations, telephone (202) 366-9826.
I. Public Participation and Request for Comments
FMCSA encourages you to participate by submitting comments and
If you submit a comment, please include the docket number for this
notice (FMCSA-2014-0177), indicate the specific section of this
document to which each comment applies, and provide a reason for each
suggestion or recommendation. You may submit your comments and material
online or by fax, mail, or hand delivery, but please use only one of
these means. FMCSA recommends that you include your name and a mailing
address, an email address, or a phone number in the body of your
document so the Agency can contact you if it has questions regarding
To submit your comment online, go to http://www.regulations.gov and
put the docket number, ``FMCSA-2014-0177'' in the ``Keyword'' box, and
click ``Search.'' When the new screen appears, click on ``Comment
Now!'' button and type your comment into the text box in the following
screen. Choose whether you are submitting your comment as an individual
or on behalf of a third party and then submit. If you submit your
comments by mail or hand delivery, submit them in an unbound format, no
larger than 8\1/2\ by 11 inches, suitable for copying and electronic
filing. If you submit comments by mail and would like to know that they
reached the facility, please enclose a stamped, self-addressed postcard
FMCSA will consider all comments and material received during the
comment period and may change this notice based on your comments.
Viewing Comments and Documents
To view comments, as well as documents mentioned in this preamble
as being available in the docket, go to http://www.regulations.gov and
insert the docket number, ``FMCSA-2014-0177'' in the ``Keyword'' box
and click ``Search.'' Next, click ``Open Docket Folder'' button and
choose the document listed to review. If you do not have access to the
Internet, you may view the docket online by visiting the Docket
Management Facility in Room W12-140 on the ground floor of the DOT West
Building, 1200 New Jersey Avenue SE., Washington, DC 20590, between 9
a.m. and 5 p.m., e.t., Monday through Friday, except Federal holidays.
The FMCSA is dedicated to reducing crashes, injuries, and
fatalities involving large trucks and buses. The Compliance, Safety,
Accountability (CSA) program is FMCSA's enforcement model that allows
the Agency and State Partners to address motor carrier safety problems
before crashes occur. The foundation of CSA is the SMS, which
quantifies the on-road safety performance of motor carriers to
prioritize enforcement resources.
The SMS uses recordable crash records involving commercial motor
vehicles (CMVs) that are submitted by the States through the Agency's
Motor Carrier Management Information System (MCMIS) to assess motor
carriers' crash risk and prioritize them for safety interventions using
the SMS Crash Indicator. To define recordable crash, the Agency relies
on the definition of ``accident'' found in 49 CFR 390.5, which means
(1) except as provided in paragraph (2) of the definition, an
occurrence involving a CMV operating on a highway in interstate or
intrastate commerce that results in: (i) A fatality; (ii) bodily injury
to a person who, as a result of the injury, immediately receives
medical treatment away from the scene of the accident; or (iii) one or
more motor vehicles incurring disabling damage as a result of the
accident, requiring the motor vehicle(s) to be transported away from
the scene by a tow truck or other motor vehicle. (2) The term accident
does not include: (i) An occurrence involving only boarding and
alighting from a stationary motor vehicle; or (ii) an occurrence
involving only the loading or unloading of cargo.
A CMV is also defined at 49 CFR 390.5, as any self-propelled or
motor vehicle used on a highway in interstate commerce to transport
passengers or property when the vehicle: (1) Has a gross vehicle weight
rating or gross combination weight rating, or gross vehicle weight or
gross combination weight, of 4,536 kg (10,001 pounds) or more,
whichever is greater; or (2) is designed or used to transport more than
eight passengers (including the driver) for compensation; or (3) is
designed or used to transport more than 15 passengers, including the
driver, and is not used to transport passengers for compensation; or
(4) is used in transporting material found by the Secretary of
Transportation to be hazardous under 49 U.S.C. 5103 and transported in
a quantity requiring placarding under regulations prescribed by the
Secretary under 49 CFR, subtitle B, chapter I, subchapter C.
Because the crash data reported to FMCSA by the States does not
specify a motor carrier's role in the crash, the Crash Indicator uses
all of a motor carrier's recordable crashes, and is not available
publicly. The Crash Indicator does weight crashes based on crash
severity, however, with more weight given to fatality and injury
crashes than to those that meet the definition of an accident only
because one or more vehicles was towed from the scene.
Research on this issue conducted by FMCSA, as well as independent
organizations, has demonstrated that crash involvement, regardless of
role in the crash, is a strong indicator of future crash risk. In fact,
the Crash Indicator is one of the strongest predictors of crashes
within the SMS. FMCSA's recently completed SMS Effectiveness Test (ET)
shows that motor carriers above the Intervention Threshold in the Crash
Indicator have a future crash rate that is 85 percent higher than the
national average (https://csa.fmcsa.dot.gov/Documents/CSMS_Effectiveness_Test_Final_Report.pdf). This document and related
reports are available in the docket of this notice.
Since FMCSA has implemented the SMS, some stakeholders have
expressed concern that the Crash Indicator may not identify the highest
risk motor carriers for intervention because it looks at all crashes
without regard to the role of the carrier in the crash. In response to
stakeholder interest and as part of the Agency's commitment to
continuous improvement, FMCSA has completed a study on the feasibility
of using a motor carrier's role in crashes as an indicator of future
crash risk. The analysis focused only on the three broad questions
below addressing the procedural issues surrounding a crash weighting
program and the feasibility of implementing such a program; it did not
focus on any other implications of the program. The three analysis
questions are separate analyses designed to inform Agency decisions.
Do PARs provide sufficient, consistent, and reliable
information to support crash weighting determinations?
Would a crash weighting determination process offer an
even stronger predictor of crash risk than overall crash involvement,
and how would crash weighting be implemented in the SMS?
Depending upon the analysis results for the questions
above, how might FMCSA manage the process for making crash weighting
determinations, including public input to the process?
The Agency's research plan was posted on the Agency's Web site at
http://csa.fmcsa.dot.gov/documents/CrashWeightingResearchPlan_7-2012.pdf on July 23, 2012. The resulting report is titled ``Crash
Weighting Analysis'' and is in the docket associated with this notice.
The draft research was peer reviewed, and the peer review
recommendations are also in the docket.
III. Summary of Analysis
The discussion below summarizes the results of the three questions
addressed in this analysis. Each question is addressed independently.
The FMCSA seeks comments on the analyses' approaches and results.
Because FMCSA does not receive PARs from the States, the Agency
created a database for analysis using 10,892 PARs obtained from two
national datasets: The National Highway Traffic Safety Administration
(NHTSA) Fatality Analysis Reporting System (FARS) and the National
Motor Vehicle Crash Causation Survey (NMVCCS).
Depending upon State procedures, most PARs do not indicate the
reason for a crash; therefore, the FMCSA employed a review process
based on the process developed for FMCSA's Large Truck Crash Causation
Study (LTCCS), particularly the methodology for assigning the
``critical event'' and the ``critical reason'' for the critical event.
This methodology focuses on pre-crash events, such as vehicle and
driver actions/movements, driver condition, and the environment at the
crash scene, to identify the circumstances leading to the crash.\1\ The
critical event is the event that immediately led to the crash and that
put the vehicle or vehicles on a course that made the crash
unavoidable. The critical reason is the immediate reason for the
critical event or the failure leading to the critical event, for
example, if a CMV driver drives too fast for the roadway type.
\1\ For details on the LTCCS methodology, go to http://www.ai.fmcsa.dot.gov/ltccs/default.asp?page=method.
The FMCSA reviewed the PARs and determined the critical event and
critical reason to identify a motor carrier's role in a crash and
assign a crash weighting for analysis purposes. In order to derive the
most robust analysis of each study question, the Agency used several
crash data sources, including PARs, the NMVCCS, and the MCMIS.
Question 1: Do PARs provide sufficient, consistent, and reliable
information to support crash weighting determinations?
One of the key questions for this study is whether FMCSA could make
reliable crash weighting determinations based solely on PARs, since the
PAR is often perceived as the most common and timely record of a crash.
This analysis (1) reviewed PAR sufficiency for determining a motor
carrier's role in a crash; (2) compared a sample of PARs with other
data sets to assess the reliability of the information on the PARs; and
(3) assessed the feasibility of identifying (coding) the motor
carrier's role for particular types of crash events without reviewing
In this study, FMCSA reviewed and coded three years of crash data,
a total of 10,892 PARs from the FARS and NMVCCS, to identify the
critical reason for the crash. Ninety-one percent of the PARs met the
criteria to be reviewed for a critical reason determination (at least
one vehicle involved in the crash was a CMV, the CMV was regulated by
FMCSA, and the crash met the criteria for a recordable crash). Nine
percent could not be reviewed because it could not be determined from
the PAR that all of these criteria were met. Of the 91 percent of the
PARs that could be reviewed, 3 percent could not be coded for a
critical reason due to incomplete, inconsistent, or insufficient
The PARs were then reviewed to determine how reliably (or
accurately) they depicted the circumstances of the crash. Specific
fields on the PARs were compared to the information in related fields
in the FARS, which provides more robust information than the PAR alone.
The FMCSA did not attempt to infer these data fields from the narrative
sections of the PAR.
The following table provides an overview of the match rate between
PARs and FARS. The Agency was unable, in this type of analysis, to
establish which record, the PAR or
FARS, was more accurate, but simply identified the fact that the two
data sources were not in agreement.
Data field PAR/FARS match PAR/FARS non-match Missing PAR data
Driver Contributing Factors...... 12.6%.................... 5.3%.................... 82.0%
First Harmful Event.............. 46.9..................... 5.6..................... 47.5
Traffic-Way Flow................. 52.4..................... 14.9.................... 32.8
Weather Conditions............... 95.7..................... 3.2..................... 1.1
Roadway Surface Conditions....... 96.7..................... 2.3..................... 1.0
The FMCSA also compared the critical reasons assigned for this
study with those assigned in matching records from the NMVCCS, which
employs a similar critical reason determination process. The analysis
found that the majority of the critical reason determinations, about 90
percent, matched between these two data sources.
The Agency also assessed the practicality of coding crashes for two
types of crash events using information available in the MCMIS as an
approach to crash weighting that would not require reviewing an actual
PAR: (1) Single-vehicle crashes deemed to be ``attributable'' to the
motor carrier; and (2) both single- and multiple-vehicle crashes with
associated post-crash inspection records indicating a pre-crash out-of-
service (OOS) condition on the CMV involved. Single-vehicle
attributable crashes are those for which the MCMIS event code
description did not indicate a collision with a pedestrian; a motor
vehicle in transport; an animal; work zone maintenance equipment; or
other/unknown movable object or ``other.'' It was hypothesized that the
critical reason for these two types of crashes would be assigned to the
CMV if the PARs were reviewed. Analysis results suggest that the coding
of single-vehicle crashes without a PAR review is feasible, but is
dependent upon accurate data as to the number of vehicles involved. For
crashes with a pre-crash OOS condition, PAR reviewers did not assign
the critical reason to the CMV in a majority of cases as they did not
consider the post-crash inspection results, but the PAR alone.
Question 2: Would a crash weighting determination process offer an even
stronger predictor of crash risk than overall crash involvement, and
how would crash weighting be implemented in the SMS?
This portion of the crash weighting analysis assumed PAR
sufficiency and reliability and looked at whether a crash weighting
methodology in the SMS Crash Indicator BASIC would provide a sharper
view of the highest risk motor carriers by identifying motor carriers
with higher future crash rates. Crash weights were derived based on (1)
the critical reason assignments for the 10,892 PARs that were reviewed;
and (2) on 671 single-vehicle attributable crashes identified in the
The Agency employed various statistical and analytical approaches
to assess crash weighting benefits. The analysis used crash data from
2009-2010 to define Crash Indicator percentiles, then tracked the
future (January 2011 to June 2012) crash rate of motor carriers above
the Intervention Threshold.
The analysis applied two approaches for modifying crash weights and
analyzed the effect of each on the crash-predictive strength of the
current Crash Indicator. The first applied higher severity weights for
crashes where the critical reason was assigned to the CMV and for
single-vehicle attributable crashes and applied lower weights for
crashes that were reviewed but not assigned to the CMV. The second
approach simply removed crashes that were reviewed but not assigned to
the CMV. Both of these approaches were applied to the same two sets of
crashes: All crashes and fatal crashes only.
Results showed that modifying the Crash Indicator by changing the
crash weights based on a motor carrier's role in a crash does not
appear to improve its ability to predict future crash rates when all
crashes are considered. Modifying the Crash Indicator to include crash
weighting improves its ability to predict future crash rates when only
fatal crashes are considered. When the crash weighting methodology was
applied, the carriers that were identified for intervention had future
crash rates that are 1.8 percent to 5.0 percent higher, when removing
crashes not assigned to the CMV during the PAR review. Fatal crashes
are, however, less than 3 percent of all crashes in the MCMIS.
Question 3: How might FMCSA manage the process for making crash
weighting determinations, including public input to the process?
The objective of this part of the analysis was to identify how a
crash weighting process might be structured and, based on this process,
estimate the resources required for both start-up and ongoing
Implementing a crash weighting effort on a national scale requires
a method for uniformly acquiring the final PARs for all or a subset of
crashes; a process and system for uniform analysis; and a method for
receiving and analyzing public input.
It must be noted that FMCSA does not currently receive PARs from
the States and that they may be difficult to obtain, due to the
requirements for secure data collection and storage, which creates a
significant, albeit unknown, cost to the Agency. The annual costs for
reviewing and coding PARs, including the acceptance of public input,
will vary depending upon the number of PARs reviewed, the number of
appeals, and the crash weighting determination process established by
the Agency. This analysis estimates potential costs of between $3.9
million and $11.2 million annually.
The analysis also provided some insight into the amount of time it
would take to make these determinations. The data provided some
indication that the timeframe for the entire crash weighting
determination process, from the submission of the crash report through
the determination process, could be so significant as to make the value
of the determination questionable for the purposes of use in the SMS,
given the 24-month analysis period used by the SMS.
IV. Request for Comments
The Agency completed the study to inform decision making concerning
the feasibility of using a motor carrier's role in crashes as an
indicator of future crash risk. Based on the information that is
provided, what steps should the Agency
take regarding crash and PAR data quality? Are there other data,
research, or related materials FMCSA should take into consideration?
Dated: January 16, 2015.
T.F. Scott Darling, III,
[FR Doc. 2015-01144 Filed 1-21-15; 8:45 am]
BILLING CODE 4910-EX-P