Pith. sign in

REVIEW 3 major objections 4 minor 72 references

From Stoplights to On-Ramps: A Comprehensive Set of Crash Rate Benchmarks for Freeway and Surface Street ADS Evaluation

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper provides the first freeway-specific crash rate benchmarks for automated driving system evaluation, finding that any-injury freeway crash rates are nearly 3.5 times higher in Atlanta than in Phoenix and that benchmark choice can b

desk verdict First freeway-specific ADS crash benchmarks from public data, but the headline geographic ratios depend on VMT denominators with unquantified bias—send to review and require uncertainty analysis. read the letter →

arxiv 2508.19425 v1 pith:R6U4FWIF submitted 2025-08-26 cs.RO

John M. Scanlon , Timothy L McMurry , Yin-Hsiu Chen , Kristofer D. Kusano , Trent Victor This is my paper · ORCID
classification cs.RO
keywords crashratebenchmarksautomateddrivingsystemsfreewaysafetysurfacestreetsvehiclemilestraveledpolice-reportedcrashesimpactassessmentpoweranalysis
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that future evaluations of automated driving systems on US freeways need road-type-specific, location-specific human crash rate baselines, and it provides the first such set for several major urban counties. Using 2023 police-reported crash data and vehicle miles traveled estimates from four states, it computes crashed-vehicle rates for freeways and surface streets, broken down by crash severity and crash type. The central finding is that freeway crash rates vary substantially by geography: any-injury-reported rates are nearly 3.5 times higher in Atlanta than in Phoenix, and no two regions had similar rates across all severity levels. The paper also shows that crash type mixes change with severity—fatal freeway crashes are disproportionately single-vehicle, pedestrian, motorcyclist, and opposite-direction collisions—so low-severity performance may not predict high-severity outcomes. It closes with a power analysis quantifying the vehicle miles needed to detect given safety improvements.

What carries the argument

The machinery is a road-type classification pipeline built on federal highway functional classes: freeways are interstates and principal arterials/other freeways/expressways; everything else is a surface street. The pipeline labels each police-reported crash as freeway or surface street by first matching road names using regular expressions and then, for roads whose classification changes along their length, confirming proximity (within 400 m) to a mapped freeway using crash latitude and longitude. VMT denominators come from state-reported estimates, supplemented by a federal highway performance database where state VMT lacked road-type breakdowns, and are scaled by national passenger-vehicl

What would settle it

Compare these benchmark rates against a crash rate computed from an independent exposure source—for example, odometer readings or GPS traces from a large sample of passenger vehicles in the same counties and year. If independently measured passenger-vehicle VMT differs systematically from the state estimates by more than the crash-rate gaps reported here (e.g., the 3.5-fold spread between Atlanta and Phoenix), the geographic comparisons would not survive.

Watch

Extended reading notes

Core claim

The paper claims that freeway crash risk is neither uniform nor reducible to a single national average. Based on public police-reported crash data and VMT estimates for 2023, it finds that any-injury-reported freeway crash rates range from 0.7 per million miles in Phoenix to 2.4 per million miles in Atlanta, a 3.5-fold geographic spread, and that fatal freeway crash rates similarly differ by about threefold. Surface street crash rates were higher than freeways at every severity level, with fatal rates 1.2 to 5.3 times higher. Because these gaps are large relative to the safety effects ADS evaluations aim to detect, using a non-local or all-roads benchmark would bias the estimated safety impa

Load-bearing premise

The state-reported and federal highway performance database estimates of vehicle miles traveled, adjusted to passenger vehicles using national proportions, correctly measure how many passenger-vehicle miles were actually driven on freeways and surface streets in each county during 2023.

Editorial extensions

If this is right

  • ADS deployments on freeways must be compared against freeway-specific baselines; using an all-road or surface-street baseline would misstate safety impact in either direction.
  • Mileage requirements differ sharply by outcome: detecting a 25% improvement in police-reported freeway crashes needs roughly 21–75 million VMT, while detecting the same improvement in fatal crashes needs billions of VMT.
  • Crash type distributions differ by severity, so a fleet that performs well in low-severity crash rates cannot be assumed to be safe for the high-severity crash types (single-vehicle, VRU, opposite-direction) that dominate fatal outcomes.
  • The same public-data methodology can be extended to additional counties and future years, providing a replicable baseline for regulators and developers.
  • Location-specific benchmarks are necessary to avoid biased safety conclusions; a Phoenix ADS compared against an Atlanta benchmark could have its fatal crash risk underestimated by more than threefold.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The geographic spread is large enough that a national benchmark is not just imprecise but directionally misleading for safety-impact claims; regulators may need to require region-matched baselines.
  • The severity-dependent crash-type mix implies ADS safety cases should report not just total crash reduction but type-specific rates, especially for the rare high-severity geometries.
  • A natural next test is to validate these VMT-based estimates against independent exposure data such as GPS-based odometer readings or insurance-based mileage, which would directly check the weakest assumption.
  • The power analysis implies that early freeway ADS fleets, with tens of millions of miles, can only statistically support claims about police-reported and injury crashes, not about fatal crashes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper compiles 2023 public police-reported crash data and VMT estimates from Arizona, California, Georgia, and Texas to construct crashed-vehicle crash rates for passenger vehicles on freeways and surface streets across six urban geographic areas. Rates are disaggregated by severity (police-reported, any-injury-reported, airbag deployment, suspected serious injury+, fatal) and crash type, and the authors derive the VMT needed to detect assumed ADS safety improvements with 80% power. The central claim is that freeway crash rates vary substantially by location—notably an any-injury-reported rate of 2.4 incidents per million miles in Atlanta versus 0.7 in Phoenix, a 3.5-fold difference—and that location-specific and road-type-specific benchmarks are therefore needed for fair ADS evaluation. The paper extends the authors' prior surface-street benchmarks and is framed as the first freeway-specific benchmark set for ADS assessment.

Significance. If the benchmark rates are reliable, the paper fills a genuine gap: existing ADS benchmarks are mostly all-road or surface-street, while multiple developers are about to deploy on freeways. Strengths include exclusive use of public data, an explicit and largely replicable road-classification and crash-typology pipeline, outcome definitions aligned with ADS regulatory reporting, and a power-analysis framework that directly supports study design. The crash-type-by-severity finding—fatal freeway crashes are disproportionately single-vehicle, VRU, and opposite-direction events—is a useful, non-obvious contribution. However, the headline geographic comparisons rest on VMT denominators of heterogeneous provenance and on rate estimates with no quantified uncertainty; until these are addressed, the central claim of large, reliable geographic variation is provisional rather than established.

major comments (3)
  1. [Results, Table 5; Discussion, 'Geographic Effects on Freeway Crash Rates'] No confidence intervals are provided for any crash rate. The fatal freeway rates are based on very small counts (roughly 4–15 crashes for the year), yet the paper presents 2.8-fold geographic ratios as stable findings. The reader cannot assess whether the Atlanta/Phoenix fatal difference, or even the any-injury 3.5-fold difference, is statistically meaningful. Please provide exact Poisson confidence intervals for all rates, report crash counts and VMT denominators alongside rates, and formally test or provide credible intervals for the geographic ratios that support the headline conclusions.
  2. [Methodology, Table 2; Limitations of the Data] Every crash rate divides by state-sourced VMT whose estimation methods differ: AZ Certified Public Miles, CA Public Road Data/HPMS, GDOT Road Mileage Reports, and TX Transit Statistics. The paper itself states that 'the variance and bias due to these sampling methods have not been studied.' Because the same VMT denominator feeds all severity levels, any systematic state- or county-level bias in VMT directly contaminates the geographic and freeway-versus-surface comparisons. In addition, the passenger-vehicle share adjustment uses FHWA VM4 state-level proportions rather than county-level truck/passenger splits. Please provide a sensitivity analysis that re-estimates headline rates under alternative VMT inputs or at least bounds the plausible denominator error, and temper the geographic-dependence conclusion until such bounds are available.
  3. [Methodology, 'Outcome Levels'; Discussion, 'Geographic Effects'] The any-injury-reported outcome is adjusted for underreporting using national NHTSA multipliers applied uniformly to all four states. The paper acknowledges state-level differences in reporting thresholds and likelihood, but no state-specific validation is performed. Since any-injury-reported rate is one of the two headline comparative outcomes (along with fatal), the claimed Atlanta/Phoenix difference could reflect differential underreporting rather than differential crash risk. Please report unadjusted and adjusted rates side by side, or apply state-specific underreporting estimates, and discuss how the geographic ranking changes.
minor comments (4)
  1. [Discussion, 'Geographic Effects on Freeway Crash Rates'; Results, Table 5] The fatal-rate numbers are inconsistent: Results/Table 5 give Atlanta 14 and Phoenix 5 per billion miles, while the Discussion gives Atlanta 15 and Phoenix 4. Please reconcile.
  2. [Methodology, 'Power Analysis Methodology'] The power-analysis formula is referenced as 'below' but does not appear in the text. Because the required VMT is a key deliverable, please include the full formula and define all terms.
  3. [Methodology, 'Road Type'] The 400 m proximity threshold to HPMS freeway segments is applied uniformly, but no sensitivity analysis is shown. This threshold may misclassify frontage-road or parallel-surface-street crashes as freeway crashes, particularly in dense urban counties. A brief sensitivity check around 200–600 m would strengthen confidence.
  4. [Figure 4] The caption says the point represents the 'minimum amount of VMT required at that performance level,' but the statistical derivation and the role of the Poisson/normal approximation should be stated explicitly in the text so readers can reproduce the calculation.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: benchmark rates are direct ratios of external police-reported crash data and state/FHWA VMT data; no fitted quantity is renamed as a prediction.

full rationale

The paper's central outputs—freeway and surface-street crash rates by geographic area and severity—are computed as direct ratios of publicly sourced police-reported crash counts to publicly sourced VMT estimates from state mileage reports and FHWA HPMS. No parameter is fitted to a subset of the outcome data and then re-presented as a predicted rate; the headline '3.5 times higher in Atlanta vs Phoenix' is an arithmetic consequence of the externally sourced crash numerators and VMT denominators. Same-author prior papers are cited for methodological conventions (outcome-level definitions from Scanlon et al. [32], crash typology from Kusano et al. [19]), but those citations supply classification schemes, not the numerical benchmark values, which are recomputed from 2023 state data. The any-injury underreporting adjustment relies on independent NHTSA estimates. The paper candidly states in 'Limitations of the Data' that 'the variance and bias due to these sampling methods have not been studied in this current study'; this is a data-quality limitation affecting robustness of geographic comparisons, not a self-referential derivation. No equation in the paper reduces to its own inputs, no uniqueness theorem is imported from the authors' prior work, and no empirical pattern is merely renamed. Thus no circular step is present.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

No parameters are fitted to a target result; the 400 m threshold is a hand-chosen classifier setting. The main burden is data quality assumptions about VMT measurement, road-type classification, and underreporting adjustment, all of which directly enter the benchmark rates.

free parameters (1)
  • Freeway proximity threshold = 400 m
    Hand-chosen radius used to match crashes to HPMS freeway segments when a road name can be both freeway and surface street. Misclassification at this step shifts crashes between the freeway and surface street numerators. Location: Methodology, Road Type.
assumptions (5)
  • domain assumption Police-reported state crash databases, with the national NHTSA underreporting adjustment for any-injury rates, provide crash counts that are comparable across the five study areas.
    Cites 60% PDO and 32% non-fatal injury underreporting and applies a national correction; cross-state comparisons of any-injury rates assume uniform reporting gaps. Location: Outcome Levels.
  • domain assumption VMT estimates from state reporting and FHWA HPMS, adjusted by national VM4 passenger-vehicle proportions, accurately represent passenger-vehicle miles by road type and county.
    This is the denominator of every rate; the authors acknowledge variance and bias from sampling methods were not studied. Location: Data Source, Limitations of the Data.
  • domain assumption Road-name regular expressions, the 400 m HPMS proximity rule, and Google geocoding for missing California coordinates assign crashes to freeways vs surface streets with negligible error.
    No validation set is provided; classification errors would directly bias the freeway benchmarks. Location: Methodology, Road Type.
  • domain assumption State KABCO injury classifications for any-injury, suspected serious injury+, and fatal outcomes are consistent enough for cross-state comparison.
    The paper claims consistency for these outcomes but warns each state has a unique KABCO definition. Location: Outcome Levels.
  • domain assumption Unknown vehicle types in crash records can be imputed from the known type distribution at the geographic level without biasing passenger-vehicle rates.
    Imputation is applied to a small proportion of vehicles whose type could not be determined; if those differ systematically, numerators are biased. Location: Data Source, In-Transport Passenger Vehicle Type.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Stoplights to On-Ramps: A Comprehensive Set of Crash Rate Benchmarks for Freeway and Surface Street ADS Evaluation." pith.science (2026). https://pith.science/paper/R6U4FWIF

@misc{pith2026250819425,
  author       = {Pith},
  title        = {Pith review of: From Stoplights to On-Ramps: A Comprehensive Set of Crash Rate Benchmarks for Freeway and Surface Street ADS Evaluation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R6U4FWIF}},
  note         = {Machine review of arXiv:2508.19425}
}
read the original abstract

This paper presents crash rate benchmarks for evaluating US-based Automated Driving Systems (ADS) for multiple urban areas. The purpose of this study was to extend prior benchmarks focused only on surface streets to additionally capture freeway crash risk for future ADS safety performance assessments. Using publicly available police-reported crash and vehicle miles traveled (VMT) data, the methodology details the isolation of in-transport passenger vehicles, road type classification, and crash typology. Key findings revealed that freeway crash rates exhibit large geographic dependence variations with any-injury-reported crash rates being nearly 3.5 times higher in Atlanta (2.4 IPMM; the highest) when compared to Phoenix (0.7 IPMM; the lowest). The results show the critical need for location-specific benchmarks to avoid biased safety evaluations and provide insights into the vehicle miles traveled (VMT) required to achieve statistical significance for various safety impact levels. The distribution of crash types depended on the outcome severity level. Higher severity outcomes (e.g., fatal crashes) had a larger proportion of single-vehicle, vulnerable road users (VRU), and opposite-direction collisions compared to lower severity (police-reported) crashes. Given heterogeneity in crash types by severity, performance in low-severity scenarios may not be predictive of high-severity outcomes. These benchmarks are additionally used to quantify at the required mileage to show statistically significant deviations from human performance. This is the first paper to generate freeway-specific benchmarks for ADS evaluation and provides a foundational framework for future ADS benchmarking by evaluators and developers.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

72 extracted references · 64 canonical work pages

  1. [1]

    Early estimate of motor vehicle traffic fatalities in 2024

    National Highway Traffic Safety Administration. "Early estimate of motor vehicle traffic fatalities in 2024." Publication DOT HS 813 710 (2025)

    work page 2024

  2. [2]

    (R) Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles,

    SAE International Surface Vehicle Recommended Practice, “(R) Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles,” SAE Standard J3016, Revised April 2021

    work page 2021

  3. [3]

    Economic and environmental benefits of automated electric vehicle ride-hailing services in New York City

    Zeng, Teng, Hongcai Zhang, Scott J. Moura, and Zuo-Jun M. Shen. "Economic and environmental benefits of automated electric vehicle ride-hailing services in New York City." Scientific Reports 14, no. 1 (2024): 4180

    work page 2024

  4. [5]

    International Organization for Standardization (ISO). 2021. Road vehicles- Prospective safety performance assessment of pre-crash technology by virtual simulation - Part 1: State-of-the-art and general method overview. ISO/TR 21934-1

    work page 2021

  5. [6]

    The Effectiveness of Electronic Stability Control (ESC) in Reducing Real Life Crashes and Injuries

    Lie, Anders, Claes Tingvall, Maria Krafft, and Anders Kullgren. 2006. “The Effectiveness of Electronic Stability Control (ESC) in Reducing Real Life Crashes and Injuries.” Traffic Injury Prevention 7 (1): 38–43. doi:10.1080/15389580500346838

    work page doi:10.1080/15389580500346838 2006

  6. [7]

    Glassbrenner D, Starnes M. 2009. Lives saved calculations for seat belts and frontal air bags (Report No. DOT HS 811 206). Washington,DC: National Highway Traffic Safety Administration

    work page 2009

  7. [8]

    Strandroth J, Rizzi M, Olai M, Lie A, Tingvall C. 2012. The effects of studded tires on fatal crashes with passenger cars and the benefits of electronic stability control (esc) in Swedish winter driving. AccidAnal Prev. 45:50–60. doi:10.1016/j.aap.2011.11.005

    work page doi:10.1016/j.aap.2011.11.005 2012

  8. [9]

    Effectiveness of low speed autonomous emergency braking in real-world rear-end crashes

    Fildes B, Keall M, Bos N, Lie A, Page Y, Pastor C, Pennisi L, Rizzi M,Thomas P, Tingvall C., 2015. Effectiveness of low speed autonomous emergency braking in real-world rear-end crashes. Accid Anal Prev.81:24–29. doi:10.1016/j.aap.2015.03.029

    doi:10.1016/j.aap.2015.03.029 2015

Show all 72 references
  1. [10]

    Isaksson-Hellman I, Lindman M. 2015. Evaluation of rear-end colli-sion avoidance technologies based on real world crash data.Proceedings of the Future Active Safety Technology Towards ZeroTraffic Accidents (FASTzero), Gothenburg, Sweden. p. 9–11

    2015

  2. [11]

    Cicchino JB. 2017. Effectiveness of forward collision warning and auton-omous emergency braking systems in reducing front-to-rear crashrates. Accid Anal Prev. 99(Pt A):142–152. doi:10.1016/j.aap.2016.11.009

    2017 doi

  3. [12]

    Cicchino JB. 2018. Effects of lane departure warning on police-reported crash rates. J Safety Res. 66:61–70. doi:10.1016/j.jsr.2018.05.006

    2018 doi

  4. [13]

    Teoh ER. 2022. Motorcycle antilock braking systems and fatal crash rates: updated results. Traffic Inj Prev. 23(4):203–207. doi:10.1080/15389588.2022.2047957

    2022

  5. [14]

    Cummings M. 2024. Assessing readiness of self-driving vehicles. In:The 103rd Transportation Research Board (TRB) Annual Meeting.Washington, D.C

    2024

  6. [15]

    Chen JJ, Shladover SE. 2024. Initial Indications of Safety of Driverless Automated Driving Systems. arXiv preprint arXiv:2403.14648

    2024 arXiv

  7. [16]

    Di Lillo L, Gode T, Zhou X, Atzei M, Chen R, Victor T. 2024. Comparative safety performance of autonomous-and human drivers: a real-world case study of the Waymo one service. Heliyon.10(14):e34379. doi:10.1016/j.heliyon.2024.e34379

    2024 doi

  8. [17]

    Di Lillo L, Gode T, Zhou X, Scanlon JM, Chen R, Victor T. 2025. Do autonomous vehicles outperform latest-generation human-driven vehicles? a comparison to waymo’s auto liability insurance claims at25.3M miles. Mountain View (CA): Waymo LLC

    2025

  9. [19]

    Comparison of Waymo Rider-Only Crash Rates by Crash Type to Human Benchmarks at 56.7 Million Miles

    Kusano, Kristofer D., John M. Scanlon, Yin-Hsiu Chen, Timothy L. McMurry, Tilia Gode, and Trent Victor. 2025. “Comparison of Waymo Rider-Only Crash Rates by Crash Type to Human Benchmarks at 56.7 Million Miles.” Traffic Injury Prevention, May, 1–13. doi:10.1080/15389588.2025.2499887

    2025

  10. [20]

    Banerjee S, Jha S, Cyriac J, Kalbarczyk ZT, Iyer RK. 2018. Hands off the wheel in autonomous vehicles?: a systems perspective on over a million miles of field data. In: 2018 48th Annual IEEE/IFIPInternational Conference on Dependable Systems and Networks(DSN), IEEE. p. 586–597...

    2018

  11. [21]

    2016.Automated vehicle crash rate comparison using naturalistic data.Final report

    Blanco M, Atwood J, Russell S, Trimble T, McClafferty J, Perez M. 2016.Automated vehicle crash rate comparison using naturalistic data.Final report. Blacksburg, VA: Virginia Tech Transportation Institute

    2016

  12. [22]

    Perez, and Julie A

    Hankey, Jonathan M., Miguel A. Perez, and Julie A. McClafferty. Description of the SHRP 2 naturalistic database and the crash, near-crash, and baseline data sets. Virginia Tech Transportation Institute, 2016

    2016

  13. [23]

    Dixit V, Chand S, Nair D. 2016. Autonomous vehicles: disengagements, accidents and reaction times. PLoS One. 11(12):e0168,054. doi:10.1371/journal.pone.0168054

    2016 doi

  14. [24]

    Favarò F, Nader N, Eurich S, Tripp M, Varadaraju N. 2017. Examiningaccident reports involving autonomous vehicles in california. PLoSOne. 12(9):e0184,952. doi:10.1371/journal.pone.0184952

    2017 doi

  15. [25]

    Schoettle B, Sivak M. 2015. A preliminary analysis of real-world crash-es involving self-driving vehicles. Report no. UMTRI-2015-34. AnnArbor, MI: University of Michigan Transportation Research Institute

    2015

  16. [26]

    Comparison of automated vehicle struck-from-behind crash rates with national rates using naturalistic data

    Goodall, Noah J. "Comparison of automated vehicle struck-from-behind crash rates with national rates using naturalistic data." Accident Analysis & Prevention 154 (2021): 106056, https://doi.org/10.1016/j.aap.2021.106056

    2021

  17. [27]

    Goodall NJ. 2021. Potential crash rate benchmarks for automated vehicles. Transp Res Rec. 2675(10):31–40. doi:10.1177/03611981211009878

    2021 doi

  18. [28]

    Teoh ER, Kidd DG. 2017. Rage against the machine? Google’s self-driving cars versus human drivers. J Safety Res. 63:57–60. doi:10.1016/j.jsr.2017.08.008

    2017 doi

  19. [29]

    Cooper, D., et al., TNCs 2020: A Profile of Ride-Hailing in California, SFCTA, 2023, San Francisco County

    2020

  20. [30]

    Flannagan C, Leslie A, Kiefer R, Bogard S, Chi-Johnston G, Freeman L,Huang R, Walsh D, Joseph A. 2023. Establishing a crash rate benchmark using large-scale naturalistic human ridehail data. Ann Arbor (MI):University of Michigan Transportation Research Institute (UMTRI)

    2023

  21. [31]

    Rage against the machine? Google's self-driving cars versus human drivers

    Teoh, Eric R., and David G. Kidd. "Rage against the machine? Google's self-driving cars versus human drivers." Journal of safety research 63 (2017): 57-60

    2017

  22. [32]

    Benchmarks for Retrospective Automated Driving System Crash Rate Analysis Using Police-Reported Crash Data

    Scanlon, John M., Kristofer D. Kusano, Laura A. Fraade-Blanar, Timothy L. McMurry, Yin-Hsiu Chen, and Trent Victor. 2024. “Benchmarks for Retrospective Automated Driving System Crash Rate Analysis Using Police-Reported Crash Data.” Traffic Injury Prevention 25 (sup1): S51–65. ...

    2024

  23. [33]

    Chen YH, Scanlon JM, Kusano KD, McMurry TL, Victor T. 2024. Dynamic benchmarks: spatial and temporal alignment for ADS performance eval-uation. arXiv preprint arXiv:2410.08903. doi:10.48550/arXiv.2410.08903

    2024 doi arXiv

  24. [34]

    Highway functional classification concepts, criteria and procedures 2023 Edition

    Federal Highway Administration (FHWA). “Highway functional classification concepts, criteria and procedures 2023 Edition” (2023). https://rosap.ntl.bts.gov/view/dot/72430/dot_72430_DS1.pdf

    2023

  25. [35]

    (2018), isbn: 978-1-56051-676-7

    American Association of State Highway and Transportation Officials, A Policy on Geometric Design of Highways and Streets, 7th Edition , American Association of State Highway and Transportation Officials: Washington, D.C. (2018), isbn: 978-1-56051-676-7

    2018

  26. [36]

    Crash and injury prevention estimates for intersection driver assistance systems in left turn across path/opposite direction crashes in the United States

    Bareiss, Max, John Scanlon, Rini Sherony, and Hampton C. Gabler. "Crash and injury prevention estimates for intersection driver assistance systems in left turn across path/opposite direction crashes in the United States." Traffic injury prevention 20, no. sup1 (2019): S133-S138

    2019

  27. [37]

    Injury mitigation estimates for an intersection driver assistance system in straight crossing path crashes in the United States

    Scanlon, John M., Rini Sherony, and Hampton C. Gabler. "Injury mitigation estimates for an intersection driver assistance system in straight crossing path crashes in the United States." Traffic injury prevention 18, no. sup1 (2017): S9-S17

    2017

  28. [38]

    Federal Highway Administration. 2023b. Highway statistics 2021. https://www.fhwa.dot.gov/policyinformation/statistics/2021/

    2021

  29. [39]

    National Center for Statistics and Analysis. 2023b. Fatality analysis reporting system analytical user’s manual, 1975-2021. Tech. Rep. DOTHS 813 417, National Highway Traffic Safety Administration

    1975

  30. [40]

    Arizona Department of Transportation. 2023a. 2022 Motor Vehicle Crash Facts for the State of Arizona. Phoenix (AZ): the Arizona Department of Transportation

    2022

  31. [41]

    Arizona Department of Transportation. 2023c. Records center. https://azdot.govqa.us/WEBAPP/_rs/supporthome.aspx

  32. [42]

    Arizona Department of Transportation. 2023b. Extent and travel dash-board. https://experience.arcgis.com/experience/ac0948fc05224aa8a80313f59a634fde

  33. [43]

    California Highway Patrol (CHP). 2024. California Crash ReportingSystem (CCRS) – Dataset – California Open Data Portal. CaliforniaHighway Patrol; 2024 May 31 [updated 2025 Feb 3]. https://data.ca.gov/dataset/ccrs

    2024

  34. [44]

    California Department of Transportation (Caltrans). 2023. California Public Road Data: statistical Information Derived from the Highway Performance Monitoring System. California Department of Transportation

    2023

  35. [45]

    Crash Reporting ,

    Georgia Department of Transportation, “Crash Reporting ,” Georgia Department of Transportation, accessed July 2025, https://www.dot.ga.gov/GDOT/pages/CrashReporting.aspx

    2025

  36. [46]

    Road & Traffic Data ,

    Georgia Department of Transportation, “Road & Traffic Data ,” Georgia Department of Transportation, accessed July 2025, https://www.dot.ga.gov/GDOT/Pages/RoadTrafficData.aspx

    2025

  37. [47]

    Texas Department of Transportation (TXDOT). 2022. Motor Vehicle Traffic Crash Data. Texas Department of Transportation [accessed 2023 Aug 15]. https://www.txdot.gov/apps-cg/crash_records/form.htm

    2022

  38. [48]

    Texas Department of Transportation (TXDOT). 2023. 2023 Texas Transit Statistics. TXDOT Public Transportation Division

    2023

  39. [49]

    Highway Performance Monitoring System Field Manual

    Federal Highway Administration. Highway Performance Monitoring System Field Manual. https://www.fhwa.dot.gov/policyinformation/hpms/fieldmanual/page01.cfm, 2018. Accessed Jul. 22,2024

    2018

  40. [50]

    Vehicle Identification Number (VIN) Requirements

    "Vehicle Identification Number (VIN) Requirements." Code of Federal Regulations, title 49, part 565 (2025). https://www.ecfr.gov/current/title-49/subtitle-B/chapter-V/part-565

    2025

  41. [51]

    2023.Framework for a conflict typology including contributing factors foruse in ADS safety evaluation

    Kusano KD, Scanlon JM, Brännström M, Engström J, Victor T. 2023.Framework for a conflict typology including contributing factors foruse in ADS safety evaluation. Proceeding of 27th InternationalTechnical Conference on the Enhanced Safety of Vehicles. PaperNumber 23-0328

    2023

  42. [52]

    Najm WG, Smith DL. 2007. Definition of a pre-crash scenario typology forvehicle safety research. Proceeding of 27th International TechnicalConference on the Enhanced Safety of Vehicles. Paper Number 07-0412

    2007

  43. [53]

    Blincoe L, Miller T, Wang JS, Swedler D, Coughlin T, Lawrence B, GuoF, Klauer S, Dingus T. 2023. The economic and societal impact of motor vehicle crashes, 2019 (revised). Tech. Rep. DOT HS 813 403, National Highway Traffic Safety Administration

    2023

  44. [54]

    Governors Highway Safety Association and National Highway Traffic Safety Administration. 2017. Mmucc guideline model minimum uniform crash criteria, 5th edition. Tech. Rep. DOT HS 812 433. Available at www.nhtsa.gov/mmucc-1

    2017

  45. [55]

    Portney, Leslie Gross, and Mary P. Watkins. Foundations of clinical research: applications to practice. Vol. 892. Upper Saddle River, NJ: Pearson/Prentice Hall, 2009

    2009

  46. [56]

    San Francisco Department of Public Health-Program on Health, Equityand Sustainability. 2017. Vision zero high injury network: 2017update – a methodology for San Francisco, California. Tech. rep

    2017

  47. [57]

    American Association of State Highway and Transportation Officials, Highway Safety Manual (Washington, D.C.: AASHTO, 2010)

    2010

  48. [58]

    Explaining the road accident risk: Weather effects

    Bergel-Hayat, Ruth, Mohammed Debbarh, Constantinos Antoniou, and George Yannis. "Explaining the road accident risk: Weather effects." Accident Analysis & Prevention 60 (2013): 456-465

    2013

  49. [59]

    Injury Risks and Crashworthiness Benefits for Females and Males: Which Differences Are Physiological?

    Brumbelow, Matthew L., and Jessica S. Jermakian. 2021. “Injury Risks and Crashworthiness Benefits for Females and Males: Which Differences Are Physiological?” Traffic Injury Prevention 23 (1): 11–16. doi:10.1080/15389588.2021.2004312

    2021

  50. [60]

    Farmer, C. M. (2019). The effects of higher speed limits on traffic fatalities in the United States, 1993–2017. Insurance Institute for Highway Safety

    2019

  51. [61]

    Effects of lowering speed limits on crash severity in Seattle

    Hu, Wen, and Jessica B. Cicchino. "Effects of lowering speed limits on crash severity in Seattle." Journal of safety research 88 (2024): 174-178

    2024

  52. [62]

    Are female drivers safer? An application of the decomposition method

    Li G, Baker SP, Langlois JA, Kelen GD. Are female drivers safer? An application of the decomposition method. Epidemiology. 1998 Jul;9(4):379-84. PMID: 9647900

    1998

  53. [63]

    (2025, January)

    National Center for Statistics and Analysis. (2025, January). Seat belt use in 2024 – overall results (Traffic Safety Facts Research Note. Report No. DOT HS 813 682). National Highway Traffic Safety Administration

    2025

  54. [64]

    Evaluation of the impacts of traffic states on crash risks on freeways

    Xu, Chengcheng, Pan Liu, Wei Wang, and Zhibin Li. "Evaluation of the impacts of traffic states on crash risks on freeways." Accident Analysis & Prevention 47 (2012): 162-171

    2012

  55. [65]

    Alcohol-impaired driving in rural and urban road environments: Effect on speeding behaviour and crash probabilities

    Yadav, Ankit Kumar, and Nagendra R. Velaga. "Alcohol-impaired driving in rural and urban road environments: Effect on speeding behaviour and crash probabilities." Accident Analysis & Prevention 140 (2020): 105512

    2020

  56. [66]

    S., & Gabler, H

    Johnson, N. S., & Gabler, H. C. (2015). Injury Outcome in Crashes with Guardrail End Terminals. Traffic Injury Prevention, 16(sup2), S103–S108. https://doi.org/10.1080/15389588.2015.1065976

    2015

  57. [67]

    Comprehensive Target Populations for Current Active Safety Systems Using National Crash Databases

    Kusano, Kristofer D., and Hampton C. Gabler. 2014. “Comprehensive Target Populations for Current Active Safety Systems Using National Crash Databases.” Traffic Injury Prevention 15 (7): 753–61. doi:10.1080/15389588.2013.871003

    2014

  58. [68]

    M., Kusano, K

    Scanlon, J. M., Kusano, K. D., & Gabler, H. C. (2016). Lane Departure Warning and Prevention Systems in the U.S. Vehicle Fleet: Influence of Roadway Characteristics on Potential Safety Benefits. Transportation Research Record, 2559(1), 17-23. https://doi.org/10.3141/2559-03 (O...

    2016 doi

  59. [69]

    Passenger and heavy vehicle collisions with pedestrians: assessment of injury mechanisms and risk

    Schubert, Angela, Stefan Babisch, John M. Scanlon, Eamon T. Campolettano, Robby Roessler, Thomas Unger, and Timothy L. McMurry. "Passenger and heavy vehicle collisions with pedestrians: assessment of injury mechanisms and risk." Accident Analysis & Prevention 190 (2023): 107139

    2023

  60. [70]

    A Mechanistic Approach to Modeling Omnidirectional Motorcyclist Injury Risk

    Schubert, Angela, Eamon T. Campolettano, John M. Scanlon, Thomas Unger, and Timothy L. McMurry. “A Mechanistic Approach to Modeling Omnidirectional Motorcyclist Injury Risk” Traffic Injury Prevention (Accepted Manuscript) (2025)

    2025

  61. [71]

    Y., DinparastDjadid, A., & Simoiu, C

    Engström, J., Liu, S. Y., DinparastDjadid, A., & Simoiu, C. (2024). Modeling road user response timing in naturalistic traffic conflicts: A surprise-based framework. Accident Analysis & Prevention, 198, 107460

    2024

  62. [72]

    Horstmann, G. (2006). Latency and duration of the action interruption in surprise. Cognition & Emotion, 20(2), 241-267

    2006

  63. [73]

    Looking for an out: Affordances, uncertainty and collision avoidance behavior of human drivers

    Johnson, Leif, Johan Engström, Aravinda Srinivasan, Ibrahim Özturk, and Gustav Markkula. "Looking for an out: Affordances, uncertainty and collision avoidance behavior of human drivers." arXiv preprint arXiv:2505.14842 (2025)

    2025

  64. [74]

    Waymo simulated driving behavior in reconstructed fatal crashes within an autonomous vehicle operating domain

    Scanlon, John M., Kristofer D. Kusano, Tom Daniel, Christopher Alderson, Alexander Ogle, and Trent Victor. "Waymo simulated driving behavior in reconstructed fatal crashes within an autonomous vehicle operating domain." Accident Analysis & Prevention 163 (2021): 106454. Append...

    2021

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.