keywords: Regression, Poisson, Negative Binomial, Road Crash, Count Data
Poisson and Negative Regression analyses are techniques used to model dependent variable that describe count data. The aim of this paper is to study the relationship between death from road crashes and, total cases of road crashes and time on Keffi-Lafia road. To achieve this, secondary data was obtained from RS 4.3 sector command, Nasarawa State on annual road traffic crash for the period of 2006 to 2015 on Keffi-Lafia road. The results from Poisson and Negative Binomial regression models revealed a positive significant relationship between total cases and Death from road crashes from 2006-2015 on Keffi-Lafia road. In addition, there is a significant negative trend of about 7.5% annual in the deaths from road crashes from 2006- 2015 on Keffi-Lafia road and there was no evidence of over dispersion, the Poisson regression model was more robust than the negative binomial regression model. This paper suggest that public enlightenment and sensitization of road users should be intensify by FRSC, as this can further reduce the rate of deaths from accident from 7.5% to 50% on Keffi-Lafia road.