Simultaneous Analysis of Global Decisions in Activity Travel Scheduling Process

Abstract

This paper presents a global characterization of activity travel scheduling decisions that uses a new data set recently collected in Valencia, Spain. As an innovative approach, bivariate probit sample selection models are used to take into account potential self-selectivity bias in the decision process. Model 1 studies decisions to realize or not realize activity travel episodes by considering whether they have been included in the planned agenda. Model 2 analyzes decisions to perform activity travel episodes as they were planned or to modify one or more of their attributes before execution by considering whether their realization had been previously decided. Random parameters are used in both models to accommodate heterogeneity effects. Location, timing, and duration of activity travel episodes are important explanatory variables in the two scheduling processes studied. Few demographic and socioeconomic variables affect the decision processes. Significant correlations between decisions included in both models are found. Policy implications of the analysis results are highlighted.