A GEV model with scale heterogeneity for investigating the role of mobility tool ownership in peak period non-work travel mode choices

The paper presents results of investigation on mode choice behaviour for peak period non-work trips (trips that are made during the peak period and are not linked to any other trips). Pure non-work trips within the peak period represent a significant portion of peak period traffic. In the case of the Greater Toronto and Hamilton Area (GTHA), the study area of this investigation, around 11 percent of peak period trips are pure non-work trips and auto driving is the dominant mode. Also more peak period non-work trips are made using the auto passenger mode than all transit modes combined. This heavy auto dependency for pure non-work trips in the peak period, when transit service is at its highest throughout the day, requires an improved understanding of such mode choice behaviour. This paper uses data from a household travel survey collected in the GTHA to investigate peak period pure non-work trip mode choice in the context of household mobility tool ownership (auto and transit pass ownership). The paper also proposes an advanced econometric modelling approach for capturing preference heterogeneity as well as scale heterogeneity (heteroskedasticity). The model involves systematic parameterization of the scale parameter of a GEV model. Empirical models highlight the superiority of the proposed model over a homoskedastic model. Empirical model also explains the influence of household mobility tool ownership on peak period pure non-work trip mode choices in terms of explaining both preference heterogeneity and scale heterogeneity.