Pith. sign in

REVIEW 4 major objections 5 minor 25 references

Unlocking Insights Addressing Alcohol Inference Mismatch through Database-Narrative Alignment

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read By aligning police crash narratives with structured alcohol flags, the paper identifies 2,767 alcohol-inference-mismatch crashes in Iowa and estimates an overall AIM rate of 24.03%.

desk verdict Useful framework with a central estimate that is not actually an underreporting rate; deserves a serious referee but needs major methodological cleanup. read the letter →

arxiv 2506.19342 v1 pith:Z5RHF5R2 submitted 2025-06-24 cs.LG cs.AIcs.CYstat.AP

Sudesh Bhagat , Raghupathi Kandiboina , Ibne Farabi Shihab , Skylar Knickerbocker , Neal Hawkins , Anuj Sharma This is my paper · ORCID
classification cs.LGcs.AIcs.CYstat.AP
keywords alcoholinferencemismatchcrashnarrationsdataqualityimpairedcrashesBERTmodelNLPmismatchedtrafficunderreporting
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

The paper tries to show that a large share of alcohol-involved crashes are hidden in plain sight: the structured database says "non-alcohol," but the officer-written narrative describes drinking, impairment, or odor. It builds a BERT-based text classifier that reads those narratives, trains it on manually corrected reports, and applies it to all 371,062 Iowa crashes from 2016 to 2022. The result is an estimated 2,767 alcohol inference mismatch (AIM) crashes, an overall AIM rate of 24.03%, with fatal and nighttime crashes less likely to be mismatched and crashes involving older drivers, trucks, pedestrians, and bicyclists more likely. If correct, this gives crash databases a practical correction layer and pinpoints counties and report types where officer training could improve data quality.

What carries the argument

The load-bearing mechanism is database-narrative alignment: a fine-tuned BERT classifier reads the free-text crash narrative, predicts whether alcohol was involved, and compares that prediction with the structured alcohol flag; when the narrative says alcohol and the flag says non-alcohol, the crash is scored as an AIM crash. BERT is a bidirectional transformer language model pretrained on large text corpora, here fine-tuned for binary alcohol classification on 8,914 manually corrected, stratified narratives. Around that core sit text preprocessing (personally identifiable information removal, punctuation and number stripping, lemmatization, TF-IDF weighting), a two-stage pipeline that filters model-predicted alcohol crashes before computing AIM rates, and a county-level analysis combining Local Moran's I spatial clustering with a probit model using random county intercepts to identify where and why mismatch concentrates.

What would settle it

Take a random sample of the 2,767 AIM crashes, or of the 11,517 model-flagged alcohol crashes, and have independent human reviewers read the original narratives plus any available hospital or toxicology records without seeing the model's label. If a large share of the 2,767 flags do not actually describe alcohol involvement, the AIM rate is largely a classifier artifact; if independent records confirm the flags, the framework's estimate holds.

Watch

Extended reading notes

Core claim

The paper's central claim is that alcohol involvement can be recovered from the officer-written narrative even when the structured database flag says "no," and that doing so at scale changes the picture of alcohol-related crashes in Iowa. Using a fine-tuned BERT text classifier, the study labels each narrative as alcohol-related or not, then compares that prediction with the ALCOHOL_REL field. Of 371,062 crashes from 2016 through 2022, the model flags about 11,517 as alcohol-related; 8,750 of those were already coded alcohol, and 2,767 were originally coded non-alcohol. Those 2,767 are the paper's AIM crashes, giving an AIM percentage of 24.03%, lower than the roughly 40% mismatch in earlier police-hospital comparisons. The mismatch is not uniform: fatal crashes have the lowest AIM rate at 16.79%, unknown-injury crashes the highest at 32.48%, and a county-random-intercept probit model finds higher mismatch odds for heavy trucks, other vehicles, drivers aged 15-24 and 65 or older, pedestrians, bicyclists, and daylight or dusk conditions.

Load-bearing premise

The load-bearing premise is that a model trained and tested on a balanced, manually corrected sample classifies the entire imbalanced crash database just as accurately as it classifies the test set, even though the paper's own full-data counts suggest real-world recall is far lower.

Editorial extensions

If this is right

  • Agencies can run the classifier on incoming crash reports and automatically flag narratives that mention alcohol but were coded non-alcohol, creating a shortlist for officer review instead of relying on manual audits.
  • The 24.03% AIM estimate implies Iowa's official alcohol-crash counts understate alcohol involvement by roughly a quarter among crashes whose narratives contain alcohol signals, which would shift the denominator for alcohol-crash rates and policy targets.
  • The severity pattern, with fatal crashes lowest at 16.79% and possible or unknown-injury crashes highest at 32.48%, suggests reporting reforms should target ambiguous, low-severity crashes rather than fatal ones.
  • Regression results point to concrete training priorities: older drivers, young drivers, heavy trucks, pedestrians, and bicyclists are the report types most likely to be mismatched, and spatially clustered High-High counties deserve the most attention.

Reading between the lines

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

  • Editorial inference: the paper's full-data application implies the classifier's real-world recall is far below its 97% test recall, since only 8,750 of the 15,282 crashes originally coded alcohol were predicted alcohol, about 57%, so the AIM count and rate would need recalibration on the imbalanced population before being used as a definitive statistic.
  • An extension the authors leave implicit: the same narrative-versus-structured-field alignment could be applied to other underreported factors such as driver distraction, drug impairment, or seatbelt use, because the mechanism does not depend on alcohol specifically.
  • A testable next step would be to link the flagged AIM crashes to hospital emergency records or toxicology results for a sample of cases; if external records agree with the narrative-based flags, the 24.03% estimate gains independent support.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper proposes a BERT-based text-classification framework to identify "alcohol inference mismatch" (AIM) in Iowa crash data, defined as crashes whose narrative mentions alcohol but whose structured ALCOHOL_REL field is non-alcohol. The authors train a BERT Large Uncased model on a manually corrected stratified sample of 8,914 narratives, report 98% test accuracy, apply it to all 371,062 crashes, and obtain 11,517 predicted-alcohol crashes, of which 2,767 are originally non-alcohol and are labeled AIM, yielding an AIM rate of 24.03%. They then fit a probit-logit model with county-level random intercepts to the machine-generated AIM outcome and perform LISA spatial clustering to identify counties with high mismatch. The paper concludes that fatal and nighttime crashes have lower AIM, while older drivers, unknown vehicle types, and vulnerable road users have higher AIM.

Significance. If the central AIM estimate were valid, the paper would provide a practical, scalable tool for improving crash-data quality and targeting enforcement or training resources. The application to a full state database is a useful proof of concept, and the integration of NLP labeling with regression and spatial analysis is a reasonable design template for similar data-quality problems. However, the manuscript does not make code or data available, and the reported validation metrics do not transfer to the full-data application, which undermines the quantitative headline claims.

major comments (4)
  1. [Analysis on the whole data; Table 5] The full-data predictions directly contradict the reported test-set performance. Table 5 reports test recall of 0.97 and precision of 0.99 for the alcohol class, but the text states that applying the model to all 371,062 crashes yields 11,517 predicted-alcohol crashes, of which 8,750 are originally alcohol-labeled. Since Table 2 sums to 15,282 originally alcohol-labeled crashes, the implied full-data recall is 8,750/15,282 = 57.2% and precision is 8,750/11,517 = 76.0%, far below the validation metrics. Because the AIM rate is computed as 2,767/11,517, both the numerator and denominator depend on the model's decision threshold at the true 4% base rate of alcohol-labeled crashes, which is not a regime represented by the balanced validation set. The paper never reports full-data precision or recall, never calibrates the threshold, and therefore the 24.03% AIM estimate is unsupported.
  2. [Preparing Training Dataset and Training the Model; Table 5] The support counts in Table 5 are internally inconsistent with the stated sample size. Training supports are 3,559 alcohol and 3,142 non-alcohol, summing to 6,701; test supports are 898 and 902, summing to 1,800; the total is 8,501, not the 8,914 stated in the text. This discrepancy suggests missing or duplicate records in the stratified sample, and it is not explained. The reported accuracy, recall, and precision are therefore computed on a sample whose composition is unclear.
  3. [Regression modeling of predicted 'Mismatch Category'; Table 4] The outcome variable "Mismatch Category" is defined solely by the BERT model's prediction combined with the original structured label (Table 4). The probit-logit model is then fitted to this machine-generated outcome, so the estimated coefficients and associated interpretations describe the classifier's disagreement patterns, not independently verified underreporting. Without a human-validated random sample of the full-data predictions (e.g., a sample of the 2,767 AIM crashes and the 8,750 non-AIM crashes), the regression and LISA spatial results cannot be interpreted as identifying factors of true alcohol reporting mismatches.
  4. [Methodology; 'AIM Percentage Estimation'] The definition of the AIM percentage is inconsistent across the paper. The Methodology first says the AIM percentage is 'the ratio of the number of AIM crashes to the total number of alcoholic crashes in a given region during a specific time period,' then later says it is computed as the number of AIM crashes divided by the total number of crashes, 'which includes both Non-AIM Crashes and those identified as AIM.' The reported 24.03% uses 2,767/11,517, i.e., the denominator is all model-positive crashes, not all alcoholic crashes (which would include 15,282 originally alcohol-labeled crashes plus an unknown number of true positives among originally non-alcohol crashes). The ambiguity materially changes the headline statistic and must be resolved.
minor comments (5)
  1. [Text Classification subsection] The paper states that TF-IDF weighted features were used to train BERT Large Uncased, but BERT is normally trained on raw token sequences, not TF-IDF-weighted vectors. Please clarify the actual input representation, as the described pipeline is not standard.
  2. [Table 5 and Figure 2] The confusion matrices in Figure 2 are referenced but not described in the text. Please include the exact cell counts so that readers can verify the reported precision, recall, and F1 values.
  3. [Table 7 and Table 8] Table 7 lists 'DriveDistracted' as a variable with 1,703 yes and 6,087 no, but Table 8 does not include this variable in the final model. Please explain whether it was excluded and why.
  4. [References] Reference 18 is a Google Drive link for 'Probit and Logit Models Example'; this is not a stable or citable source. Please replace it with a proper textbook or peer-reviewed citation for probit/logit modeling.
  5. [Throughout] There are several typographical errors and inconsistent terms, e.g., 'Unkown' in Table 2, 'Probit Logit' vs 'probit logistic regression', and 'alcohal' in the abstract. A careful proofread is needed.

Circularity Check

2 steps flagged · score 6.0 of 10

AIM rate and regression outcomes are constructed from the BERT classifier's own predictions, so the central estimates describe the model's false-positive behavior rather than independently verified underreporting.

  1. fitted input called prediction [Section 'Analysis on the whole data and AIM Percentage Estimation']
    "The filtered dataset comprised approximately 11,517 records, with 2,767 categorized as "AIM crashes" and 8,750 as "Non-AIM crashes". Consequently, the overall AIM percentage, or AIM rate, was estimated to be 24.03%."

    By the counts in the same paragraph, 2,767 is the number of crashes the BERT model predicted as alcoholic while the original field was non-alcohol, and 8,750 is the number predicted alcoholic with original field alcohol. Therefore 24.03% = 2,767/(2,767+8,750) is exactly the false-discovery proportion of the fitted BERT classifier on the full data, i.e. 1 minus full-data precision. No BAC, hospital, or expert re-review is introduced to verify that these model positives are true underreported alcohol crashes, so the headline estimate is a direct summary of the fitted model's outputs rather than an independent measurement.

  2. self definitional [Section 'Regression modeling of predicted "Mismatch Category" of each crash using Probit Logit' and Table 4]
    "The labeled data for the "Mismatch Category" for each crash, along with other relevant crash variables outlined in Table 1, were used to build the Probit Logit model."

    Table 4 defines "Mismatch Category" from the original ALCOHOL_REL label and the BERT-predicted label. Since the predicted label is generated by the same fitted BERT model, the probit/logit outcome is a deterministic function of that model's predictions. The coefficients (older drivers, bicyclists, daylight, unknown severity, etc.) therefore describe where the classifier disagrees with the original labels, not independently established causes of underreporting. The paper's own limitation, "NLP-based output must be validated by subject matter experts," is never addressed with an external benchmark, making the regression findings reducible to properties of the fitted classifier.

full rationale

The paper's formal derivation chain does not rely on self-citations or imported uniqueness theorems, and there is no external mathematical claim that forces the result. The circularity is partial and operational: the BERT model is fitted to a manually corrected subsample, and then the same model's outputs are used to define the AIM outcome and the AIM rate. The full-data estimate 24.03% equals the classifier's false-discovery share (2,767/11,517) under the original labels, and the probit/logit regression is fitted to that machine-generated outcome, so the identified "factors" describe the classifier's disagreement pattern rather than externally verified underreporting. The paper's own limitation notes that NLP output must be validated by subject matter experts, but no such validation is reported, and the full-data precision and recall implied by the paper's own counts (about 76% and 57%) are far from the balanced test-set metrics. This reduces the central claims to properties of the fitted model rather than independently established alcohol-inference mismatch. However, this is not a formal self-citation or a definitional equivalence of the entire derivation, because the fitted model could in principle be a valid instrument; hence the score is 6 rather than 8-10.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central claim rests on three unreviewed assumptions: narratives preserve alcohol evidence, manual labels are gold, and test accuracy transfers to the full imbalanced population. No external validation is provided.

free parameters (1)
  • BERT classification probability threshold = 0.5 (default)
    The AIM count and rate are directly determined by the cutoff used to map BERT's probability output to a binary alcohol label. The paper does not state or vary this threshold.
assumptions (3)
  • domain assumption Crash narratives contain explicit, recoverable linguistic evidence of alcohol involvement when alcohol was present.
    The whole AIM detection rests on narratives containing such evidence; if officers omit alcohol mentions in narratives, the model cannot find them.
  • domain assumption The manually corrected labels in the 8,914-crash training sample are accurate gold-standard labels.
    No inter-annotator agreement, annotation protocol, or independent verification (e.g., BAC tests) is reported; the model's objective function is defined by these labels (Section 'Preparing Training Dataset').
  • domain assumption Classifier performance measured on the balanced test set transfers to the full 371,062-crash population with its natural class imbalance.
    The test set is 50/50 alcohol/non-alcohol, while the full data has roughly 4% alcohol; the paper applies the model directly without recalibration (Section 'Analysis on the whole data').

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unlocking Insights Addressing Alcohol Inference Mismatch through Database-Narrative Alignment." pith.science (2026). https://pith.science/paper/Z5RHF5R2

@misc{pith2026250619342,
  author       = {Pith},
  title        = {Pith review of: Unlocking Insights Addressing Alcohol Inference Mismatch through Database-Narrative Alignment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Z5RHF5R2}},
  note         = {Machine review of arXiv:2506.19342}
}
read the original abstract

Road traffic crashes are a significant global cause of fatalities, emphasizing the urgent need for accurate crash data to enhance prevention strategies and inform policy development. This study addresses the challenge of alcohol inference mismatch (AIM) by employing database narrative alignment to identify AIM in crash data. A framework was developed to improve data quality in crash management systems and reduce the percentage of AIM crashes. Utilizing the BERT model, the analysis of 371,062 crash records from Iowa (2016-2022) revealed 2,767 AIM incidents, resulting in an overall AIM percentage of 24.03%. Statistical tools, including the Probit Logit model, were used to explore the crash characteristics affecting AIM patterns. The findings indicate that alcohol-related fatal crashes and nighttime incidents have a lower percentage of the mismatch, while crashes involving unknown vehicle types and older drivers are more susceptible to mismatch. The geospatial cluster as part of this study can identify the regions which have an increased need for education and training. These insights highlight the necessity for targeted training programs and data management teams to improve the accuracy of crash reporting and support evidence-based policymaking.

Figures

Figures reproduced from arXiv: 2506.19342 by the authors.

Figure 1
Figure 1. Methodology Analysis Objective 1: Framework for identifying AIM crashes using BERT Text Classification BERT Text Classification The study employed the BERT Large Uncased model, a bidirectional, unsupervised language representation model known for its capability to handle complex linguistic structures (Janstrup et.al., 2016). The framework for achieving this objective is illustrated in [PITH_FULL_IMAGE:figures/full_… view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 23 canonical work pages

  1. [1]

    Under-reporting of motor vehicle traffic crash victims in New Zealand, Accident Analysis and Prevention

    Alsop, J., Langley, J., 2001. Under-reporting of motor vehicle traffic crash victims in New Zealand, Accident Analysis and Prevention

    work page 2001

  2. [2]

    Under-reporting of road crash casualties in France

    Amoros, E., Martin, J.L., Laumon, B., 2006. Under-reporting of road crash casualties in France. Accid Anal Prev 38 4 , 627–635. doi:10.1016/j.aap.2005.11.006

    work page doi:10.1016/j.aap.2005.11.006 2006

  3. [3]

    A Large Language Model Framework to Uncover Underreporting in Traffic Crashes

    Arteaga, C., Park, J., 2025. A Large Language Model Framework to Uncover Underreporting in Traffic Crashes

    work page 2025

  4. [4]

    Exemption of behind-the-wheel driving test for novice young drivers: a serious public health concern

    Bari, I., Paichadze, N., Hyder, A.A., 2020. Exemption of behind-the-wheel driving test for novice young drivers: a serious public health concern. Journal of road safety 31 4 , 59–61. doi:10.33492/JRS-D-20-00252

    work page doi:10.33492/jrs-d-20-00252 2020

  5. [5]

    February 2023 6

    Blincoe, L., Miller, T.R., Wang, J.-S., Swedler, D., Coughlin, T., Lawrence, B., Guo, F., Klauer, S., Dingus, T., 2023. February 2023 6. Performing Organization Code 7. Authors 13. Type of Report and Period Covered NHTSA Technical Report 14. Sponsoring Agency Code Unclassified

    work page 2023

  6. [6]

    MedlinePlus

    Blood Alcohol Level: MedlinePlus Medical Test [WWW Document], 2007. . MedlinePlus. URL https://medlineplus.gov/lab-tests/blood-alcohol-level/ (accessed 7.25.24)

    work page 2007

  7. [7]

    FARS Encyclopedia: Help - Terms [WWW Document]

    Bozak David J, 1996. FARS Encyclopedia: Help - Terms [WWW Document]. NHTSA. URL https://www-fars.nhtsa.dot.gov/help/terms.aspx (accessed 1.12.25)

    work page 1996

  8. [8]

    State Variation in Underreporting of Alcohol Involvement on Death Certifi cates: Motor Vehicle Traffi c Crash Fatalities as an Example, Stud

    Castle, I.P., Hingson, R.W., White, A.M., 2014. State Variation in Underreporting of Alcohol Involvement on Death Certifi cates: Motor Vehicle Traffi c Crash Fatalities as an Example, Stud. Alcohol Drugs

    work page 2014

Show all 25 references
  1. [9]

    Beta regression in R

    Cribari-Neto, F., Zeileis, A., 2010. Beta regression in R. J Stat Softw 34 2 , 1–24. doi:10.18637/JSS.V034.I02

    2010 doi

  2. [10]

    Under-reporting of road traffic injuries to the police: Results from two data sources in urban India

    Dandona, R., Kumar, G.A., Ameer, M.A., Reddy, G.B., Dandona, L., 2008. Under-reporting of road traffic injuries to the police: Results from two data sources in urban India. Injury Prevention 14 6 , 360–365. doi:10.1136/ip.2008.019638 Bhagat et al. 19

    2008 arXiv

  3. [11]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

    Devlin, J., Chang, M.-W., Lee, K., Google, K.T., Language, A.I., 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

    2019

  4. [12]

    URL https://www.nhtsa.gov/risky-driving/drunk-driving#the-issue-alcohol-effects (accessed 4.24.24)

    Drunk Driving | Statistics and Resources | NHTSA [WWW Document], 2022. URL https://www.nhtsa.gov/risky-driving/drunk-driving#the-issue-alcohol-effects (accessed 4.24.24)

    2022

  5. [13]

    Incomplete Accident Reporting: Meta-Analysis of Studies Made in 13 Countries | Enhanced Reader

    Elvik Rune, Mysen Borger Anne, 1999. Incomplete Accident Reporting: Meta-Analysis of Studies Made in 13 Countries | Enhanced Reader. Transp Res Rec 1665

    1999

  6. [14]

    Improving the performance of aspect based sentiment analysis using fine-tuned Bert Base Uncased model

    Geetha, M.P., Karthika Renuka, D., 2021. Improving the performance of aspect based sentiment analysis using fine-tuned Bert Base Uncased model. doi:10.1016/j.ijin.2021.06.005

    2021 doi

  7. [15]

    Understanding traffic crash under-reporting: Linking police and medical records to individual and crash characteristics

    Janstrup, K.H., Kaplan, S., Hels, T., Lauritsen, J., Prato, C.G., 2016. Understanding traffic crash under-reporting: Linking police and medical records to individual and crash characteristics. Traffic Inj Prev 17 6 , 580–584. doi:10.1080/15389588.2015.1128533

    2016

  8. [16]

    Police narrative reports: Do they provide end- users with the data they need to help prevent bicycle crashes? Accid Anal Prev 164

    Lopez, D., Malloy, L.C., Arcoleo, K., 2022. Police narrative reports: Do they provide end- users with the data they need to help prevent bicycle crashes? Accid Anal Prev 164. doi:10.1016/j.aap.2021.106475

    2022

  9. [17]

    Underreporting of Driver Alcohol Involvement in United States Police and Hospital Records: Capture-Recapture Estimates

    Miller, T.R., Gibson, R., Zaloshnja, E., Blincoe, L.J., Kindelberger, J., Strashny, A., Thomas, A., Ho, S., Bauer, M., Sperry, S., Peng, J., Singleton, M., Smith, T.J., Zhang, Y., 2012. Underreporting of Driver Alcohol Involvement in United States Police and Hospital Records: ...

    2012

  10. [18]

    Probit and Logit Models Example.pdf - Google Drive [WWW Document]

    Katchova Ani, 2013. Probit and Logit Models Example.pdf - Google Drive [WWW Document]. URL https://docs.google.com/file/d/0BwogTI8d6EEieGV4Rk9xQi1aVzA/edit?resourcekey=0- cCrdpXty6l5e5kOT0lZgqA (accessed 4.25.24)

    2013

  11. [19]

    World Health Organization

    Road Traffic Injuries [WWW Document], 2023. . World Health Organization. URL https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries (accessed 7.21.24)

    2023

  12. [20]

    Salifu, M., Teresa, M., 2009. Road safety data: collection and analysis for target setting and monitoring performances and progress UNDER-REPORTING OF ROAD TRAFFIC CRASH DATA IN GHANA Update and improvement of the traffic accident data collection procedures in Spain: The METRA...

    2009

  13. [21]

    Injured drivers and alcohol use: culpability, convictions, and pre- and post-crash driving history - PubMed [WWW Document]

    Soderstrom, C.A., Birschbach, J.M., Dischinger, P.C., 1990. Injured drivers and alcohol use: culpability, convictions, and pre- and post-crash driving history - PubMed [WWW Document]. J Trauma. URL https://pubmed.ncbi.nlm.nih.gov/2213929/ (accessed 7.25.24)

    1990

  14. [22]

    Recording of alcohol in official crash statistics: Underreporting and procedures to improve statistics

    Vissers, L., Houwing, S., Wegman, F., 2018. Recording of alcohol in official crash statistics: Underreporting and procedures to improve statistics. J Australas Coll Road Saf 29 3 , 15–22

    2018

  15. [23]

    Investigation of winter weather crash injury severity using winter storm classification techniques

    Walker, C.L., Farooq, M.U., Cecava, J., 2024. Investigation of winter weather crash injury severity using winter storm classification techniques. doi:10.1016/j.trip.2024.101073

    2024

  16. [24]

    Estimating under-reporting of road crash injuries to police using multiple linked data collections

    Watson, A., Watson, B., Vallmuur, K., 2015. Estimating under-reporting of road crash injuries to police using multiple linked data collections. Accid Anal Prev 83, 18–25. doi:10.1016/J.AAP.2015.06.011

    2015 doi

  17. [25]

    A method to account for and estimate underreporting in crash frequency research

    Wood, J.S., Donnell, E.T., Fariss, C.J., 2016. A method to account for and estimate underreporting in crash frequency research. Accid Anal Prev 95, 57–66. doi:10.1016/j.aap.2016.06.013

    2016 doi

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.