Student travel behavioural preferences in Mahikeng: evidence

University student mobility is not reflected in the National Learner Transport Policy although some formal operators make provisions for identifiable post-school learners. South African universities do not accommodate the majority of students and they tend to have scattered campuses and residences. Only a few (6/26) public universities have contracts with scheduled bus and shuttle services specifically for students. Literature reveals that the characteristics of university student mobility are distinct from the general population. A segment specific approach to redress the potentially problematic results of aggregation could guide the treatment and inclusion of post-matric mobility needs in the National Learner Transport Policy.

This study presents evidence of behavioural heterogeneity in the context of university
student travel behaviour. It fills a policy and research gap by exploring university student travel behaviour and making a unique contribution to stated choice literature and applications in Africa.

Two hypotheses are tested. First, students have unique compositions of behaviour influencing their intention to use bus and minibus taxi. Secondly, there are level of service (LOS) preference differences between students who have high, medium or low intent to use any public transport mode. In navigating toward these hypotheses, the Theory of Planned Behaviour is used to theoretically reflect student behavioural inclinations toward bus and minibus taxi services in Mahikeng. In order to represent the choices students make between two modes the Hybrid Discrete Choice Modelling framework is adopted and applied. Therefore the hypotheses mentioned above are tested by means of grouping student responses in terms of the likelihood that they belong a certain level of intention to use a mode, namely: high (P), neutral (N) or low (Z). And a behaviour specific latent class choice model (LCCM) is developed to estimate the probability of a student choosing a specific mode. Utilities are estimated in the form of multinomial logit models that are group (class) specific.

An unlabelled d-optimal survey is developed based on observation and literature. Distributed at the North West University’s Mahikeng site of delivery, the survey had 121 properly completed responses of 150 printed copies. Most of the respondents were female (65%); males were older; generally they hope to own a car 2 years after graduation and 86% say they prefer using minibus taxi. The LCCM results are twofold: class membership and class specific choice, as shown in the table below. Most of the sample have a low intention to use bus (63%), and high intentions to use minibus taxi (17.58%).

The class specific choice results reveal that high intention students prefer minibus taxi LOS, whilst low and neutral (indifferent) students prefer bus LOS. This evidence is further compounded by a joint model offering behaviour specific mode choice probabilities. In terms of LOS, the value of certain attributes is also estimated. It is found that the high intention class derives a benefit from seating availability; the low intention class has a low value of travel time and waiting time is inflated in the non-class specific model.

After the model estimations, scenario tests are conducted. The scenarios reveal that waiting time and travel time related utility functions are volatile and sensitive changing the market shares dramatically. The scenarios provide encouraging evidence that behaviour specific LOS designs are plausible, at least theoretically.

One major finding is that segmenting markets based on behavioural theory provides useful insight about travel behaviour—outside of demographic variables. It is shown that different behavioural groups, have different behavioural inclinations. Voluntary and regulatory triggers aimed at changing behaviour may not be pivoted to influence the most important variables related to a certain mode. From a policy perspective this study presents opportunities for travel behaviour change research and service contracts with class specific subsidies, for example. In terms of LOS design, the results intimate that LOS offering can be distributed behaviourally in order to attract choice and lifestyle users of public transport. Specific additions pertaining to university student mobility issues are directed at the National Learner Transport Policy and other legislation. Through this approach it may become increasingly appropriate to argue that LOS performance contracts should engage travel behaviour change related initiatives (i.e. marketing campaign, free ticket day). Further work is needed to practically connect preference with policy interventions that account for price, quality, and behavioural constraints. This will be particularly important in the age of mobility applications and other mobile technology in transport industries.


A working paper encompassing parts of this study was presented at the International Choice Modelling Conference in April 2017 with the title ‘Estimating Student Travel Preferences in Mahikeng: A Latent Class Approach Based on Behavioural Indicators‘. This was sponsored by the PTV Group which featured us in a short article ‘From pedal cycling in Kenya to bus demand in Johannesburg’. Thereafter, an updated version of the working paper was presented at the Southern African Transport Conference in July of the same year (although not listed in the program).

My studies at the University of Cape Town were funded by the Southern Centre of Development and then later by the Education, Training and Development Practices Sector Education and Training Authority (ETDP SETA). The project has been submitted for examination (6/2/2018), once the final copy is available a link will be embedded.


Creative Commons License


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s