Advancing the Behavioral Safety Analytic Tools Capabilities of the Connecticut Department of Transportation

Wang, Kai; Shaon, Mohammad Razaur Rahman; Shirani, Niloufar; Tucker, Andrew; Russell, Dan; Jackson, Eric; Chaudhary, Neil; Tison, Julie · 2021 · ROSA P / Connecticut. Dept. of Transportation

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses the limitations of existing highway safety analytics, which primarily rely on roadway geometry and traffic variables, by developing tools specifically for driver behavior-related crashes. Since driver behavior contributes to over 90% of crashes, yet traditional models like the Highway Safety Manual (HSM) lack behavioral emphasis areas, the Connecticut Department of Transportation (CTDOT) sought to create a comprehensive analytic framework. The project aimed to develop crash prediction models for six specific behavioral emphasis areas—impaired driving, aggressive driving, young drivers, motorcycles, pedestrians, and distracted driving—and to build a web-based application to assist practitioners in hotspot identification and countermeasure selection. The methodology involved collecting and integrating multidisciplinary data at the town level for the period of 2015–2019. Data sources included crash records, roadway geometry, traffic volume, crime and citation records, toxicology results, socioeconomic demographics, and business location data. To address multicollinearity in high-dimensional datasets, Principal Component Analysis (PCA) was applied to crime/citation and toxicology variables. Crash prediction models were then estimated using Negative Binomial (NB) regression to account for over-dispersion in crash counts. The models predicted total crashes, as well as crashes with severe (K+A) and moderate (B+C) injuries. Additionally, the researchers compiled a novel behavioral countermeasure database by synthesizing information from sources such as the CDC and NCHRP, extracting Crash Modification Factors (CMFs) and assigning quality ratings to 377 identified countermeasures. The analytical results demonstrated that the first five principal components for citation data and three for toxicology data accounted for over 95% of the variance in their respective original variables. Eighteen NB regression models were successfully estimated, providing coefficient estimates for the six behavioral crash types across different injury severity levels. These models identified significant risk factors associated with driver behaviors, utilizing the integrated unconventional data elements. The resulting countermeasure database provided detailed descriptions, costs, implementation times, and effectiveness ratings for behavioral interventions, filling a gap left by engineering-focused resources like the CMF Clearinghouse. The significance of this work lies in the development of the Connecticut Behavioral Safety Tool (CTBST), a customized web application that operationalizes these findings. The tool enables practitioners to perform the full highway safety management process, including network screening, diagnosis, countermeasure selection, and economic appraisal. By incorporating non-traditional data and focusing on behavioral factors, the CTBST provides a more holistic approach to safety management, allowing agencies to better identify hotspots and select effective, behavior-specific countermeasures to mitigate crashes.

Key finding

The study produced negative binomial crash prediction models for six driver behavior-related crash types using integrated multidisciplinary data and developed a web-based application tool that combines these models with a comprehensive behavioral countermeasure database to support highway safety management processes.

Methodology

modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• incidence prevalence
• induced exposure
• demographic disparities
• causation analyses
• comparative international
• naturalistic crash near crash

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: crash risk outcomes, observational prevalence
• Methodological Resource: dataset resource