Public Comments

Status: Original and derived artifacts are available for this document.

Download: pdf

March 29, 2026
National Highway Traffic Safety Administration
Docket Management Facility
U.S. Department of Transportation
1200 New Jersey Avenue SE
Washington, DC 20590
Re: Public Comment on NHTSA’s Agency Information Collection Activities; Notice and
Request for Comment; Crash Report Sampling System, Non-Traffic Surveillance, and
Special Study Data Collection (Docket No. NHTSA– 2026–0463; OMB Control Number
2127–0714)
I am responding to the National Highway Traffic Safety Administration’s (NHTSA) March 19,
2026 notice requesting comment on the proposed extension with modification of the Crash
Report Sampling System (CRSS) information collection. I support continuation of CRSS as a
national crash data resource. However, NHTSA should revise this Information Collection
Request (ICR) submission to address a specific and longstanding flaw in the data on which
CRSS depends. Underride crashes are underreported in State police crash reporting systems, and
those missed cases bias the national underride injury estimates derived from CRSS. [1]
This is a practical utility issue under the Paperwork Reduction Act (PRA). The current notice
expressly requests comment on ways to enhance the quality, utility, and clarity of the
information to be collected. [1] That standard is directly implicated here. CRSS is used to
estimate the overall crash picture and to support cost-benefit analyses of highway safety
initiatives and regulations. [2] Because CRSS is a probability sample of police-reported crashes,
systemic underride underreporting in State crash reports produces a downward-biased national
underride injury estimate. If underride is not identified in the source report, the case may still
enter the CRSS sample, but it is counted as an ordinary crash rather than as an underride crash.
Sampling weights cannot correct for crashes that were never identified as underride in the
underlying records. [1][2][3][4][5]
I have diligently raised this same underride data issue with the agency for years through other
formal channels. In April 2023, I submitted an Administrative Procedure Act petition asking
USDOT and NHTSA to standardize underride definitions, differentiate rear underride, side
underride, and front override, require an underride or override checkbox on State police crash
report forms as States modernize electronic reporting, and provide recurring training to analysts
and law enforcement. In September 2023, I submitted a public comment on NHTSA’s State Data
Transfer for Vehicle Crash Information ICR and again tied underride capture to State electronic
data collection, MMUCC alignment, and standardized data transfer to NHTSA. NHTSA has
substantial practical leverage to influence how States collect and report crash data that populate
CRSS. [6][7][9]
The following points explain why NHTSA should revise this ICR submission to address the
underride data deficiencies identified below.

2
1. Underreported underride data is not a new issue.
NHTSA analyses and Government Accountability Office (GAO) findings have documented that
underride crashes are underreported because reporting practices vary across jurisdictions and
many police crash report forms do not include an underride field. [3][4][5] The GAO concluded
that inconsistent definitions and the absence of a dedicated underride data element can affect
whether underride crashes are captured in national crash databases. [3] The GAO also found that
only 17 police crash report forms had a specific field for underride. [3] That same undercounting
problem also affects injury estimation when CRSS relies on the same underlying State crash
reporting systems.
2. Underreported underride affects the accuracy and utility of CRSS data and associated
rulemaking.
CRSS is used to estimate national crash injuries and to support regulatory analysis. [2] NHTSA’s
side underride ANPRM explains that the agency had to review crash data and estimate sideunderride underreporting through police crash report review to define the injury target population
more accurately. [4] If underride is not identified in the State crash reporting systems from which
CRSS draws its sample, national injury estimates for underride crashes will be understated, and
the benefits of countermeasures designed to prevent those injuries will likewise be understated.
NHTSA should improve underride recording when the original crash report is completed and
transferred through State crash data systems, rather than continuing to rely on indirect
adjustments after the fact. [1][2][4][5]
3. NHTSA’s April 23, 2025 response does not resolve the underlying data collection
problem.
NHTSA’s April 23, 2025 response to prior correspondence states that the agency cannot require
States to provide specific data via the FARS program, cannot mandate modifications to State
police crash report forms through FARS, and does not treat participation in FARS as mandatory.
[9] That response does not resolve the known data collection issue. NHTSA already has
authorities and incentives it can use outside the FARS program, including Section 405(c) trafficrecords funding, crash-data modernization support, MMUCC guidance, State Data Transfer and
other electronic reporting initiatives, training, and strategic-plan oversight. Those tools can be
used to press for adoption of a dedicated underride checkbox on police crash report forms that
supply CRSS source data. [1][7][8][9] Moreover, Section 402 requires each State to have a
highway safety program approved by the Secretary, requires those programs to comply with
uniform guidelines promulgated by the Secretary, and requires satisfactory assurances including
participation in FARS and development of statewide data systems to provide timely and effective
data analysis. [6]
The traffic records grant program provides a direct mechanism to address this problem. Under 23
U.S.C. § 405(c), these grants support State programs that improve the timeliness, accuracy,
completeness, uniformity, integration, and accessibility of State safety data. The implementing
regulation, 23 CFR 1300.22, requires a State seeking a traffic records grant to have a traffic
records strategic plan, approved by the TRCC and updated annually, describing specific,

3
quantifiable, and measurable improvements in its core safety databases, including the crash
database, and to demonstrate quantitative improvement in attributes such as accuracy or
completeness. [6] That matters here because underride undercounting is precisely an accuracy
and completeness problem in a core crash database.
Appendix A to Part 1300 also confirms that NHTSA is not limited to passive encouragement. It
provides that grant recipients must comply with other reporting, data collection, and evaluation
requirements as prescribed by law or detailed in program guidance. [6] That language undercuts
any suggestion that NHTSA lacks a meaningful basis to use grant administration, program
guidance, and traffic-records oversight to press for specific data-quality expectations.
The issue is whether NHTSA is using the broader authorities and leverage Congress has already
provided to obtain State adoption of a dedicated underride checkbox or data element in police
crash report forms and associated crash data systems, rather than continuing to rely on indirect
adjustments after the data reach CRSS. [6][7][8]
4. Underride guidance and a pamphlet are not a substitute for a discrete data element and
consistent State capture.
NHTSA’s underride crash pamphlet may address part of the GAO recommendation for training
and guidance, but it is not a substitute for a dedicated underride checkbox or field on the police
crash report form. [3][5][9] Unless underride is recorded as a discrete data element in the State
crash report and carried through State crash data systems into CRSS, the national estimate will
continue to undercount underride crashes. That shortcoming also matters under DOT Order
1351.29A, which states that DOT should maximize the utility of the information collected, that
information collections must be of practical utility, and that information collection instrument
designs should be consistent with the intended use of the information. [10]
5. NHTSA has practical tools to address this data collection and reporting problem.
I offer this recommendation based not only on policy analysis but also on extensive professional
experience administering Federal-State cooperative agreements tied to data collection and
performance requirements. During a 30-year career with the U.S. Fish and Wildlife Service, I
managed and oversaw cooperative agreements and grant programs with States under Section 6 of
the Endangered Species Act. These agreements provided Federal funding to support State data
collection and reporting activities and required measurable progress toward agreed objectives. In
practice, Federal agencies routinely use funding conditions, strategic planning requirements, and
performance review processes to influence how States collect and report data. NHTSA’s trafficrecords grants, crash-data modernization programs, and cooperative agreements supporting
national crash databases operate in a functionally similar manner. [7]
6. Passenger-vehicle and vulnerable road user underride crashes should be tracked
separately.
Vulnerable road user underride crashes should be identified and reported separately from
passenger-vehicle underride crashes. CRSS and FARS include pedestrian and bicyclist cases, but

4
NHTSA’s 2023 side underride analysis focused on passenger vehicles sliding under the side of
trailers and did not evaluate vulnerable road user fatalities in that analysis. [4] These events
differ materially from passenger-vehicle underride scenarios. To remain analytically consistent
and useful for safety and cost-benefit analyses, NHTSA should track vulnerable road user
underride crashes through separate data elements and explicit coding manual guidance and
related instructions. [1][4][5]
Requested action
NHTSA has long recognized that underride cases are missed in national crash databases and has
relied on proxy adjustments to compensate for that limitation. The agency administers the
programs needed to improve the data at the source and should use those tools now. [6][7] Each
of the recommendations in this comment is directed at improving the quality, clarity, and utility
of the information collected under this ICR, consistent with the PRA’s purpose of ensuring that
federal data collections are necessary, accurate, and useful. These requested changes would also
better align this ICR with DOT Order 1351.29A, including its direction that DOT information
collections be of practical utility and be designed consistently with their intended use. [10]
Therefore, NHTSA should revise this ICR submission and related data-collection practices to
correct the underride data deficiencies in the State crash reporting systems that supply CRSS
source data. Specifically, NHTSA should:
1. Use its Section 405(c) traffic-records grant authority, Section 402 highway safety program
oversight, traffic-records strategic-plan requirements, crash-data modernization programs, and
program guidance to press States to adopt a dedicated underride data element or checkbox in
police crash report forms and associated crash data systems.
2. Establish concrete underride-reporting adoption timelines, completeness benchmarks, and
related corrective actions in federally supported traffic-records plans, electronic crash-reporting
initiatives, and related grant administration.
3. Revise underride data elements, coding manuals, and reporting guidance so that pedestrian,
bicyclist, and other vulnerable road user events involving going under the side or rear of a truck
or trailer are explicitly identified and tracked separately from passenger-vehicle underride
crashes.
4. Take corrective action to improve the accuracy of CRSS underride data capture at the source
to align the information collection of OMB Control No. 2127-0714 with DOT Order 1351.29A.
5. Report publicly, or at minimum in the administrative record for this ICR, on the actions taken,
the timelines and benchmarks established, and the measurable progress made toward adoption of
underride reporting in the State crash reporting systems that supply CRSS source data.
For transparency and to ensure meaningful public participation, NHTSA should also clearly
identify any subsequent Office of Management and Budget review-stage notice for this ICR,

5
including where and how interested parties may submit comments during Office of Information
and Regulatory Affairs review. [10][11]
Sincerely,
Eric Hein
[1] National Highway Traffic Safety Administration. March 19, 2026. Agency Information
Collection Activities; Notice and Request for Comment; Crash Report Sampling System (CRSS),
Non-Traffic Surveillance (NTS), and Special Study Data Collection. [Docket No. NHTSA-20260463], 91 FR 13397-13402.
[2] National Highway Traffic Safety Administration, Crash Report Sampling System, stating that
CRSS is used to estimate the overall crash picture and forms the basis for cost and benefit
analyses of highway safety initiatives and regulations.
[3] Government Accountability Office. 2019. Truck underride guards: Improved data collection,
inspections, and research needed. GAO-19-264.
[4] National Highway Traffic Safety Administration. April 21, 2023. Advance notice of proposed
rulemaking; side underride guards. [Docket No. NHTSA–2023–0012]; 88 FR 24535-24543.
[5] National Highway Traffic Safety Administration. 2022. Data collection and reporting of
underride crashes by law enforcement. 2 pp.
[6] National Highway Traffic Safety Administration. February 6, 2023. Uniform Procedures for
State Highway Safety Grant Programs. [Docket No. NHTSA-2022-0036], 88 FR 7780-7832.
[7] National Highway Traffic Safety Administration. May 3, 2023. Agency Information
Collection Activities; Notice and Request for Comment; State Data Transfer for Vehicle Crash
Information. [Docket No. NHTSA-2023-0019], 88 FR 27949-27954.
[8] National Highway Traffic Safety Administration (2025, February, Revised). MMUCC
guideline: Model minimum uniform crash criteria, 6th edition (Report No. DOT HS 813 525a).
[9] Chen, C.-L. April 23, 2025. Letter to Eric Hein regarding underride/override data in the
Fatality Analysis Reporting System (FARS). National Highway Traffic Safety Administration. 3
pp. (enclosed)
[10] U.S. Department of Transportation. December 8, 2022. DOT Order 1351.29A, Managing
Information Collections Under the Paperwork Reduction Act.
[11] Clearance of collections of information, other than those contained in proposed rules or in
current rules. 5 C.F.R. § 1320.10.