Exploring Factors Influencing Speeding on Rural Roads: A Multivariable Approach

Ferko, Marija; Pirdavani, Ali; Babić, Dario; Babić, Darko · 2024 · Crossref

DOI: 10.3390/infrastructures9120222

archive: archived pipeline: cataloged verified

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

Summary

This study investigates the complex dynamics of speeding on rural roads, addressing the critical need to identify factors contributing to speed limit violations to improve road safety. Speeding is a primary cause of severe road crashes, with 55% of EU fatalities occurring on rural non-motorway roads. While previous research has utilized behavioral surveys, naturalistic driving data, and simulators, there is a notable gap in studies focusing on spot speed measurements in rural environments, where unique infrastructural and traffic conditions complicate speed management. The research aims to link location characteristics, vehicle types, and traffic flow metrics with non-compliance rates, specifically examining variables such as time of day, average summer daily traffic (ASDT), roadside characteristics, and settlement location. The methodology employed a multivariable analysis framework using extensive data collected from 20 traffic counters installed on 15 distinct secondary state roads in Croatia. Data were gathered over two years during July to minimize weather-related variability, resulting in a filtered sample of 4,623,852 unique vehicle records. The study focused on tangent road sections to capture free-flow speeds, ensuring drivers were not influenced by enforcement cameras or immediate geometric constraints like curves. The dependent variable was defined as a binary classification of "Speeding" (exceeding the posted limit by >0 km/h) versus non-speeding. Independent variables included posted speed limits (ranging from 50 to 90 km/h), ASDT, roadway width, distance to the nearest intersection, roadside state, and settlement location. The analysis utilized both conventional binary logistic regression and advanced machine learning algorithms, performed using SPSS and R software, to determine the best-fitting models for predicting speeding behavior. The results indicate that all tested models achieved accuracy rates above 74%, though they exhibited higher sensitivity for predicting positive speeding cases than specificity for negative cases. The analysis identified four primary factors significantly influencing speeding across the models: the posted speed limit, the distance to the nearest intersection, roadway width, and traffic load. These findings highlight specific relationships between infrastructure design, traffic volume, and driver compliance. For instance, the proximity of intersections and the volume of traffic were found to be critical determinants of speed choice, alongside the statutory limits and physical road dimensions. The significance of this research lies in its provision of actionable insights for policymakers and law enforcement agencies. By identifying the key variables that predict speeding, the study enables the targeted development of road safety measures. It allows authorities to determine locations where speeding is most likely to occur and plan specific interventions, such as infrastructure modifications or enhanced enforcement, to reduce the frequency of speeding vehicles. This approach supports the Safe System strategy by addressing unsafe speeds through data-driven, location-specific management rather than generalized policies, ultimately aiming to reduce crash severity and fatalities on rural networks.

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.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

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

• speed choice