Analysis of Monthly Road Accidents Using Instrumental Variable

Hamza Mansur and Okafor U.L

Keywords: Road Accidents, Instrumental Variable (IV), Reduced Variable (RV), Prone zone, Regression Analysis


The issue of road accidents is an increasing problem in developing economy especially Nigeria. The number of deaths on our roads are increasing geometrically to higher percentage, thus regulating traffic movement on roads is an important task. The objective of the research is to make comparisons between two different forecasting techniques and suggest the better one. We used method of regression analysis to generate the results of the two different models (IV and RV). Among the two models IV has the highest coefficient of determination of 0.7959 and lowest standard error of estimate of 0.2041compared to RV with 0.5447and 0.4553 respectively, the results are compared using a graphical method proposed by Wilson (2015). He suggested plotting the individual fitted values against the actual observed values, if the point lies approximately along the line y = x, then the model is significant and recommended for accidents analysis.