Approaches to Modelling
RAND Europe's modelling team provides a range of quantitative expertise to help and support clients in making policy choices and operational decisions. RAND's approach is objective, scientific, takes a holistic view of the system, uses appropriately sophisticated models and methodologies, and is grounded in the analysis of relevant data sources.
Quantitative Tools: Discrete Choice Modelling
Through its professional work, the modelling team has made major innovations and extensions to best practice in the area of discrete choice modelling, as a means by which to understand and predict choice behaviour. This work is most frequently undertaken in the transport sector, but increasingly we are applying our expertise to other sectors, for example in health care, communications and provision of consumer services. Our contribution has been to take these methods out of the university environment and make them applicable to forecasting and evaluation issues in policy analysis. In particular, we have pioneered forecasting techniques and the use of mixed data sources to support modelling.
Initially this work was based primarily on data describing real choices made by people, referred to as revealed preference (RP) data, in which RAND Europe built up substantial capability. Subsequently, the use of stated preference (SP) data became equally important. Simply stated, SP analysis, sometimes called conjoint analysis, is a technique for establishing the relative importance of different attributes (characteristics or features) of a good or service. A number of procedures for eliciting responses exist, but modern practice is focussing on the use of Discrete Choice Experiments (DCE), which involve the presentation of hypothetical choice experiments to survey respondents, where each alternative in the experiment is described by relevant attributes, e.g. quality of the service, cost of the service, future characteristics, etc. Each of the attributes in the experiment is described by a number of levels, e.g. low cost, high cost. The attribute levels are combined using principles of experimental design to define different packages of goods. SP data is particularly useful for evaluation of future policies, for which no RP data on the impacts to users exists. RAND Europe was one of the first companies to employ SP DCE in the transport sector and continues to conduct research to improve SP methods. RAND Europe pioneered procedures to combine RP and SP data to exploit the strengths of each of the data types to best advantage. We also offer expertise in RP and SP survey design, based on insights gathered from our extensive practice in RP and SP modelling.
Transport Modelling
RAND Europe has contributed to the development and testing of new approaches in transport studies for over 20 years. The application of discrete choice (disaggregate) models in transport has allowed significant improvements in the quality and range of applicability of urban, regional and national (multi-modal) travel demand forecasting techniques. These improvements have enabled the development of model systems, which can accurately predict traveller responses under a wide range of policies and exogenous changes, including:
- choice of mode
- choice of route
- choice of travel destination
- choice of departure time
- car ownership and licence holding
- trip frequency
- residential location
- employer location
The last two choices are key to modelling changes in land use resulting from transport policy. The model systems we develop are typically highly detailed, containing in the order of 1000 geographical zones, multiple modes and purposes, and detailed segmentation of traveller behaviour by person and household types. Considerable attention has been given to the implications of excess demand in the form of congestion, and the resulting feedback to other stages in the demand formation process, e.g. time-of-day of travel, and ultimately the mode, destination and frequency of the trip. The modelling team frequently work with a range of different types of data, and are adept at developing procedures to overcome data bias during analysis. We have extensive experience in applying discrete choice models as part of an appraisal procedure, using the Department for Transport's Economic Efficiency of the Transport System (TEE) or other approaches.
Valuation Studies
RAND Europe has carried out value-of-time studies for national governments in a number of countries, including the UK, France, Denmark and the Netherlands. These studies have been conducted using SP techniques to assess travellers' willingness to pay for travel time savings. These valuations are a key input to the estimation of consumer surplus resulting from transport projects, which governments use to choose between competing investments. These techniques have also been applied to the analysis of new infrastructure investments, such as the construction of tolled tunnels.
Reliability of travel time is an area of increasing policy concern in many Western European countries, where congested transport networks result in unreliable journey times, so that travellers have to allow extra time for their journey in case they are delayed en-route. RAND Europe is currently undertaking a range of research in this topical area, including studies looking at the cost of unreliability to travellers, measuring reliability, and techniques to enable future levels of reliability to be forecast.
Finally, valuation studies supporting key areas of policy in other fields are also being carried out, such as communications, where we have studied the impact on business of postal delays, the value of second and third generation mobile phones and digital television, and the willingness of rural communities to pay for their local post office.
Accuracy of Outputs
Many of the techniques employed by the modelling team provide numerical outputs of interest to policy makers, such as willingness to pay for a service and predictions of traffic flows on a link. The degree of uncertainty associated with a given output gives policy makers additional insight into the viability of a given policy option. The modelling team have developed techniques to enable the calculation of errors and uncertainties in forecasting, which can account for the errors and correlations associated with parameter estimates, errors in the data, and the effects of sample sizes.
Top