The state transportation agencies (STAs) in the United States are mandated by federal rule to carry out work-zone impact assessment for highway rehabilitation projects. The work zone impact assessment requires calculating road user costs (RUCs) which is the sum of vehicle operating costs, accident costs, and value of time (VOT). The term 'value of time' refers to monetary equivalent of travel time wasted due to rehabilitation projects. In current practice, STAs assume VOT as homogeneous within their respective states. This leads to inaccurate RUCs calculations and poses many misapplications. Research has found that VOT is influenced by socio-demographic variables which vary within the states. But there is a lack of framework to evaluate the extent to which these factors affect value of time. The major objective of this research is to develop and validate a model that predicts value of time heterogeneously. The data were collected to cover 20 major cities in California. The state of California was chosen for this study because most highway rehabilitation projects are carried out there. The data sources included the United States Census Bureau, the California Department of Transportation (Caltrans), and the Bureau of Labor Statistics. With these data, a predictive model was developed using multiple linear regression analysis. Lastly, the model was validated using PRESS statistic. The results reveal that age, annual average daily traffic, and effective hourly income were the most significant factors influencing value of time. This study developed a model which will help Caltrans in calculating value of time heterogeneously and therefore, improve the accuracy of RUCs calculations. Moreover, this research will serve as a guideline for other STAs to develop models for respective states. Therefore, this model has a potential to greatly improve the accuracy of value of time and therefore, RUCs. The future research should focus on the identified factors, especially cost-of-living index and annual average daily traffic. Further research is required to account for heterogeneity due to other factors such as vehicle occupancy, frequency of travel, and educational qualifications.
