A spatio-temporal accessibility measure for modelling activity participation in discretionary activities

Accessibility is a key indicator for the number of opportunities experienced by an individual, and it is generally assumed that higher levels of accessibility lead to higher levels of activity participation and satisfaction. Despite the fact that this relationship has been tested in preceding studies, no clear pattern of correlation has been found between an individual's accessibility and his/her participation in out-of-home discretionary activities. Previous studies mostly applied aggregate, spatial accessibility measures that do not account for the heterogeneity related to an individual's spatio-temporal constraints. However, in addition to locations' spatial distribution, personal constraints related to the mandatory activities a person undertakes - and therefore, his/her available time budget throughout the day - impact activity choices and scheduling. This paper proposes a disaggregate, person-based accessibility measure that takes spatial and temporal constraints into consideration for the travel behavior of 11,599 individuals in the Wasatch Front region, Utah. This method is compared to an aggregate, home-based accessibility measure to assess both methods' ability to predict activity participation. The results show a highly significant, moderate correlation between the proposed person-based accessibility measure and participants' surveyed partaking in discretionary activities. Both the model prediction and the locally weighted regression smoother indicate that the greatest change in participation occurs within the mid-range of accessibility levels, and not in the low end of the accessibility range as was expected. In addition, the home-based method shows a negative and highly significant relationship to activity participation, which indicates that aggregate accessibility measures may provide counterintuitive findings.