Analyzing and Modeling the Effects of Independent Variables on Texas Severe Work Zone Crashes

Previous literature and research on work zone crashes seem to reach a similar conclusion that frequency and severity of crashes in work zones is dependent on causative factors that range from Endogenous factors (which in this study will be referred to as factors that are components of the work zone design) to exogenous factors (which in this study will be referred to as factors that are outside the work zone design components). While numerous studies have been undertaken on these factors, their individual influences on the resulting work zone crashes is yet to be given exhaustive analysis. Using statistical analysis and physical observations, this document will be modeling the different factors to determine their contribution to severe crash risk in work zones.

For this research, crash severity will be selected as the dependent variable. Though all crashes are worthy of investigation, their significance is measured by the cost of the damages that resulted from the crashes. In this work, high crash severity refers to crashes that resulted in either Possible injury, Non-incapacitating injury, incapacitating injury or killed. By utilizing the Multinomial Regression as the modeling technique with crash severity as the dependent variable, the significance of the various parameters in terms of their contributions to work zone crash severity is represented as P value. This measures the confidence limit of the different variables to a 95% significance level.

The importance of this finding is that most of the significant factors to severe work zone crashes are independent endogenous variables. These are variables that are likely to be corrected by proper work zone design and efficient temporal traffic planning.