Evaluation of the modal choice behaviour and bus service preferences of commuters of the scheduled Golden Arrow Bus Services (Pty) Ltd using stated choice data

Master Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

Evaluation Problem: The Western Cape Provincial Government faces a public problem of declining service levels with respect of public bus transport services. Stemming from the public problem is the management problem of modelling choice behaviour of commuter stated choices for utility maximisation and therefore as a means of optimising the allocation of the Public Transport Operating Grant (PTOG) expenditure. Historically, differing perceptions amongst travellers, and the difficulties in quantifying these attributes, mean that they are rarely included (directly) within the modelling and appraisal process, or the associated utility computation (Crockett, Sinclair and Whelan. 2008:11). A combination of policies which would ensure that the discrete choices of commuters for an improved bus service are considered in a modal shift from the Golden Arrow Bus Services (Pty) Ltd (GABS) bus service to the MyCiTi Integrated Rapid Transit (IRT) service is required to be produced from this evaluation. Evaluation Approach: The evaluation brings together data and analytical information on a broader, more inclusive, societal model of the public transport service in Cape Town. It seeks to: - identify the value aspects of the public bus transport infrastructure and services, - identify and model choice behaviour of GABS bus service users, - generate preference valuations for public transport service attributes and - determine the effect of service attributes on modal choice namely GABS and a hypothetical MyCiTi service area. A discrete choice experiment models the stated choices of respondents who were made to choose between various combinations of service levels during the morning and afternoon peak commuting times. The commuters' willingness to pay for a switch to a new MyCiTi IRT-type service from their current choice of an existing GABS bus service is estimated. The choice data is collected by means of an on-board bus survey along three particular routes in Table View, a West Coast Region in the City of Cape Town, South Africa and where the new MyCiTi service is being introduced. The analyses will show which service attributes are significant in commuter mode choice behaviour, such as changes travel time, fare prices and other significant service attributes, as well as which level of service would maximise utility for the target population. Major Evaluative Conclusions: The evaluation found that the DCE choice modelling approach used was unfamiliar to the respondents and would most likely not have been completely understood. Although the factorial approach to designing the experiment could identify an exhaustive list of value aspects to choose from, the need to adopt a fractional factorial in the final design does necessitate further experimentation to produce a more comprehensive choice model, inclusive of more service attributes and with the discrete choice models corroborated with revealed preference data. Seat availability was by far the most significant choice determinant and the lack thereof would be a serious deterrent to a modal shift to using the new MyCiTi service. The number of transfers during the trip and the distance of the bus stop from home were also significant choice determinants. The choice models indicated that the female commuters particularly, were willing to pay for the new MyCiTi bus-type service. In the absence of suitable seating capacity, a reduced travel time would be required to reduce the standing times and make the MyCiTi service an attractive option. The analysis produced inconclusive data for ridership predictions, although it can generally be said that provided sufficient seating, the GABS bus users will be willing to switch modes, as there is no indication in the data to suggest otherwise. Considering the preferences expressed for the service attributes, a hypothetical service can be proposed, with a service mix of R9.00 per trip, that would take 45 minutes and that offered the commuter a seat for the journey. The first bus stop would be no more than two kilometres away and the journey would consist of no more than one transfer to reach the final destination.