Enhancing safety: Evaluating law enforcement impact in South Carolina work zones through advanced split–plot design with blocking

This paper investigates the effectiveness of law enforcement on speed reduction in South Carolina work zones using a split–plot design with blocking, which offers improved precision over traditional factorial designs. Using 2019 data, eight models were evaluated to explore average speed responses across two key segments: the entire work zone (the full length of the construction area) and the transition area (the upstream section where drivers begin slowing down and where troopers are typically stationed). The analysis includes two variations: (1) “excess speed”, defined as the amount by which average vehicle speed exceeds the temporary posted speed limit, and (2) traffic volume, incorporated as a covariate. Results show that the presence of troopers significantly reduces both average speeds and excess speeds, particularly in the transition area. Seasonal effects were generally insignificant, although interaction effects emerged in Models 6 and 8, where traffic volume was included as a covariate. Including traffic volume slightly improved model precision. Based on these findings, the study recommends maintaining consistent law enforcement presence in work zones, particularly during winter, to maximize speed reduction. Targeted enforcement in transition areas, where officers are clearly visible, can further enhance driver compliance. The split–plot design with covariate adjustment is suggested as a robust framework for future evaluations of enforcement strategies under varying traffic and environmental conditions. • Developed an advanced split-plot design with blocking to improve accuracy in measuring speed reduction in work zones. • Police presence reduced speeds across both work zones and transition areas. • Police presence increased compliance with work zone speed limits. • Seasonal effects were analyzed, with enforcement presence having the greatest impact during winter months. • Traffic volume was incorporated as a covariate to enhance model precision and control for confounding effects.