A Mixed Outcome Model of Multimodal Safety and Mobility Outcomes

This study proposes a multivariate mixed outcome model to simultaneously examine the safety, mobility, and level of traffic stress (LTS) for vehicles and active transportation. This multimodal-multivariate model aims to gain insights into the feedback loops between active transportation safety and mobility metrics and vehiclebased safety and mobility performance metrics by employing the bi-directional causal (endogenous) interactions among the outcomes. The study used a system of endogenous equations with the account for the correlation between error terms in the equations. Each equation in that system of equations represents a performance measure used to examine safety and mobility for the different transportation modes of travel. The study used data gathered from the interstate ramp terminal intersections in Washinton state, primarily focusing on routes I-5, I-82, I-90, I-182, I-205, and I-405. The study employed the conditional mixed process (CMP) model, which helps in modeling system of equations with mixed outcomes. The results show high correlations among the heterogeneous residuals, which ensures the need to model these measures simultaneously. Additionally, it shows a direct impact from one performance measure into the other and a direct effect from one mode of travel into the other