Probability Modelling of Road Traffic Accidents on Federal Highways in Kaduna State

Road traffic accidents on federal highways in Kaduna State, North West Nigeria, represent a significant cause of preventable mortality and economic loss, with the Federal Road Safety Commission recording over 1,400 crash incidents in the state between 2020 and 2023. This study applies probability distributions, specifically the Poisson, negative binomial, and zero-inflated Poisson models, to analyse the frequency and spatial clustering of road traffic crashes on the Kaduna-Abuja and Kaduna-Zaria federal highways. Monthly crash frequency data are obtained from the FRSC Kaduna State Sector Command, categorised by crash type, time of day, vehicle category, and contributing factor. Model selection is based on the Akaike Information Criterion, with the negative binomial model selected as the best fit due to observed overdispersion in crash counts. Spatial clustering analysis using the nearest-neighbour index identifies six statistically significant accident black spots along the two corridors. The probability of a fatal crash occurring on any given day is estimated at 0.73 on the Kaduna-Abuja stretch. Regression analysis of contributing factors identifies speeding, night driving, and tyre failure as the three most significant predictors of fatal outcomes. The study recommends targeted enforcement deployment at identified black spots and installation of speed-activated warning systems. Keywords: traffic accident modelling, Poisson distribution, negative binomial, Kaduna State, road safety.