Research published in "Transport Policy" examines the factors influencing driver’s license acquisition among young people. The study focuses on Millennials and Generation Z in the United States, utilizing data-driven insights through eXtreme Gradient Boosting (XGBoost) and SHapley Additive Explanations (SHAP) to identify key socioeconomic determinants.
The study authors, Kailai Wang, Jonas De Vos, Michael Smart, and Sicheng Wang, highlight consistent predictors across both generations. These predictors include the percentage of licensed family members, household income per capita, educational attainment, and public transit ridership. Findings show that a higher number of licensed family members, particularly beyond a 0.75 threshold, significantly increases the likelihood of youth obtaining their driver’s license. Similarly, individuals with some college education or an associate degree are more likely to have a license.
The research also discovered that household income is positively associated with licensing within certain levels but declines at the highest income brackets. The authors note, "Family played an important role in youth driver’s license acquisition," challenging common assumptions that higher income directly correlates with higher licensing rates.
The study provides a detailed predictive model of licensing behavior and suggests that these insights could guide planning and policy strategies to promote equitable driver education access and sustainable mobility solutions.
The full article, "Explaining Youth Driver Licensing Determinants Using XGBoost and SHAP," is available in the journal "Transport Policy," Volume 168, Pages 87-100, 2025.
Citation:
Kailai Wang, Jonas De Vos, Michael Smart, Sicheng Wang. Explaining Youth Driver Licensing Determinants Using XGBoost and SHAP, Transport Policy, Volume 168, 2025, Pages 87-100, ISSN 0967-070X, https://doi.org/10.1016/j.tranpol.2025.04.009
