A Research Framework for Studying Transit Bus Driver Distraction

D’Souza, Kelwyn A.; Maheshwari, Sharad K. · 2013 · ROSA P / Hampton University. Eastern Seaboard Intermodal Transportation Applications Center

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 paper presents a modular research framework designed to standardize the study of transit bus driver distraction, addressing a gap in literature where such studies are typically conducted independently and inefficiently. The research was motivated by the increasing prevalence of electronic distractions and the unique nature of transit driving, where distractions often stem from factors beyond the driver’s control, such as passenger interactions and operational communications. Unlike personal vehicles, transit accidents involve higher potential for mass injury, yet distraction causes are poorly understood due to underreporting and a lack of established methodologies. The primary objective was to develop a reusable, standardized framework comprising modules for data collection, analysis, validation, and interpretation, allowing transit agencies to conduct studies efficiently across varying cost and time constraints. The methodology involved testing the framework at two transit agencies in Virginia: Hampton Roads Transit (HRT), a regional agency, and the Potomac and Rappahannock Transportation Commission (PRTC), an urban agency. Data collection utilized three sources: historical accident databases from police reports, self-administered driver perception surveys, and route observations. The survey instrument, adapted from prior ergonomic studies, collected data on driver attributes (age, gender, experience), driving patterns, and perceived distraction sources. Accident data were classified into preventable and non-preventable categories, with an estimated 17% of total accidents attributed to distraction. Analysis included Exploratory Data Analysis (EDA) to identify patterns and Confirmatory Data Analysis (CDA) using statistical modeling. A Distraction Risk Index (DRI) was developed to classify distracting activities into four risk zones based on severity. Key findings from the illustrative application of the framework revealed distinct patterns in accident frequency and driver demographics. Accident data indicated that preventable accidents, and thus distraction-related incidents, were more frequent in the Southside region of Hampton Roads compared to the Northside, with the highest occurrence on Fridays and between 12:00 PM and 6:00 PM. Driver surveys showed a mean age of 47 years, with most drivers falling within the 36–55 age range, a demographic less prone to distraction-related crashes than younger drivers. The framework successfully categorized distracting activities into risk zones and utilized multinomial logistic regression to identify significant factors impacting distraction levels. Validation methods included expert verification and simulation techniques to confirm model fit and results accuracy. The significance of this work lies in providing transit agencies with a structured, modular toolset to investigate driver distraction without reinventing methodologies for each study. By standardizing processes for data collection, risk classification, and statistical analysis, the framework enables agencies to identify specific risk factors, such as route characteristics or driver attributes, and implement targeted countermeasures. The paper concludes with guidelines for interpreting results and recommendations for improving model accuracy, such as handling missing data and outliers. This framework facilitates a more systematic understanding of transit bus driver distraction, potentially leading to reduced accident rates and improved safety for passengers and other road users.

Key finding

The research framework provides transit agencies with a standardized, modular set of methodologies for collecting distraction data, classifying activities into risk zones, and validating statistical models to predict distraction risks.

Methodology

mixed_methods

Sample size: 130

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.

• visual
• distraction detection algorithms
• external distraction
• temporal
• manual
• bus coach

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: observational prevalence
• Theoretical Contribution: conceptual framework, computational model