|
Abstract
|
Key Findings
• The negative binomial regression model was found to be the best fitting model or the model that better agrees with the data.
• Pedestrian behaviour factors of walking on roads, crossing outside pedestrian crossings, crossing carelessly, and other forms of negligence demonstrated a positive relationship with road-related fatalities.
• Behaviour factors of being under the influence of alcohol demonstrated zero influence and crossing at pedestrian crossings exhibited a small negative influence.
• There was a 1% increase in the number of crash deaths for every additional fatal crash involving pedestrians walking on a road.
Pedestrian behaviour is one of the major contributors to road fatalities. The negative binomial regression model was found to better agree with road fatality data, and this study used this model to assess the influence of pedestrian behaviour factors on road fatalities in Malawi. The data used in this analysis were crash reports of pedestrian behaviour factors and observed fatalities for the period 2000-2015 obtained from the national database, except for the 2013 data, which were disregarded because they appeared to be incomplete. Whereas pedestrian behaviour factors of walking on roads, crossing outside pedestrian crossings, and other negligent and careless behaviours were found to be positively correlated with road deaths, indicating that road-related fatalities increased with increasing input data, factors of being under the influence of alcohol and crossing at pedestrian crossings demonstrated negligible influence. The study also found that there was a 1% increase in the number of crash deaths for every additional fatal crash involving pedestrians walking on roads. Moreover, an additional 0.5% increase in the number of fatalities was recorded for every fatal crash involving a pedestrian behaviour factor of crossing outside the pedestrian crossing or other negligent behaviour. An increase of 0.3% in the number of the fatalities was seen for every extra fatal crash caused by crossing carelessly or factors other than pedestrian behaviour. Despite coefficient values being small in all variables, which is a major limitation of this study, enforcement can prioritise those variables that increase road-related fatalities or even couple them with other risk factors such as speed.
Keywords
Road Fatalities, Pedestrian Behaviour Factors, Count Data, Poisson Regression Models, Malawi
Language: en |