Evaluating Safety Training's Role in Reducing Construction Accidents: Poisson and Weibull Model Analysis

Authors : Aydın Oğuz, Osman Hansu

Pages : 133-156

Doi:10.33720/kisgd.1627168

View : 113 | Download : 145

Publication Date : 2025-12-31

Article Type : Research Paper

Abstract :This study examines how safety training affects both the frequency of workplace accidents and the duration of accident-free intervals in high-risk construction settings. Based on a 36-month dataset from six major firms, workers were divided into four groups according to training type: specialized, general, informal, or none. Using a combined methodological approach—Poisson regression, multivariate Poisson modeling, and Weibull survival analysis—the study quantifies the impact of training on both the likelihood and timing of accidents. Specialized training was associated with up to a 50% reduction in accident frequency and nearly doubled the length of safe intervals (k < 1), particularly in high-rise construction projects. General training showed moderate effects, while informal or absent training had minimal impact, with hazard rates remaining stable or rising. Drawing on Heinrich’s Safety Pyramid and Reason’s Swiss Cheese Model, and validated through AIC, BIC, and likelihood ratio tests, the findings suggest that structured safety programs lead to meaningful, lasting behavioral change. The study advocates for task-specific, technology-enhanced training—such as virtual reality and wearables—as essential components of modern safety strategies in complex construction environments.
Keywords : İnşaat Güvenliği, Güvenlik Eğitimi, Kaza Azaltma, İstatistiksel Modelleme, Sağkalım Analizi

ORIGINAL ARTICLE URL