Modelling road fatalities from tricycle crashes in Ashanti Region, Ghana: An application of regression with ARIMA errors

Abstract

Tricycles contribute to road crashes globally, especially in Asia where it is a popular means of transport. Safety assessment of tricycles has discovered that associated fatality rate and severity of injuries from tricycle crashes, which has risen steadily, are due to their flawed designs especially the absence of seatbelts and paddings of hard surfaces. Tricycle related crashes were sourced weekly from the regional office of the National Road Safety Authority from January 2012 to June 2022 and modelled with Autoregressive Integrated Moving Average Model. The results showed ARIMA (1,0,3) errors were retained as the best model for estimating road fatalities caused by tricycle crashes. The exogenous factors: tricycle crashes, time of day, season of the year, and road design were tested and found to be statistically significant relative to its contribution to the fatalities. The time series components were deemed to be relevant contributing factors to fatalities for the study period. Further, crash data forecasted to October 2022 largely exhibited a sinusoidal behaviour with spikes at various weeks. It was recommended that tricycle drivers undergo mandatory refresher courses on safe driving techniques and general road safety regulations as a policy intervention. In addition, the licensing regime for tricycles should be strengthened, and government should create separate lanes along some major routes within the metropolitan, municipal, and district assemblies to eliminate or reduce the negative incidents.

Keywords

• Tricycle
• Crashes
• Fatality
• Regression
• ARIMA errors