Much attention has been recently paid to the fact that travellers respond to the level of unreliability, and to the existence of economic benefits or costs associated with this response. This has been acknowledged in the academic literature; in the Eddington Report; in Towards a Sustainable Transport System; in early work on the NATA Refresh; and in some new WebTAG guidance on the treatment of reliability in modelling and forecasting. Nevertheless, the study of how unreliability affects travellers’ choices is still a relatively new area of transport research and an even newer area for choice modelling practitioners. This paper does not describe new research work but it is (to the best of our knowledge) the first comprehensive review of what we already know about the process of revealing the way unreliability affects travel choices, as well as what is still unknown.

We raise four main questions, the answers to which provide a broad perspective on the current capability to incorporate travellers’ attitudes to unreliability into the common tools of transport assessment. The four questions covered are the following. First, we discuss which travel responses are affected by unreliability. We demonstrate that there are questions to be asked about the response hierarchy we normally use in a full demand model. Second, we examine what data can be used to study the attitudes to unreliability. Although the dominance of SP-based evidence on the response to unreliability is unavoidable, its implications deserve careful discussion. Third, we explore what is known on the variables used in the utility function in models that consider the aversion to unreliability. We explain the difference between using attributes that travellers respond to and using attributes that capture this response empirically. Fourth, we discuss additional inputs required to study the response to unreliability in future scenarios. This relates to the quality of the match between the variables used in the demand model and the data available from a supply models. We conclude with a summary of insights which have already been established and issues that need to be investigated further.

Researchers will find in our review some insights on the most important avenues to explore when further investigating how to account for unreliability in choice models. Practitioners can use our findings to ensure that their choice models are fit for the purpose of illustrating how unreliability truly affects travel demand.