Meta-analysis of UK values of travel time An update

Numerous travel demand studies have been carried out over the past five decades many of which produce estimates of the value of travel time This includes a rich body of largely unpublished evidence which can provide valuable insights into the impact of variables such as GDP travel distance purpose and mode on this critical parameter for transport modelling and appraisal The work reported in this paper updates and extends our previous meta-analyses of UK values of time (Wardman 1998 2001a 2004) by adding recent studies and widening the range of explanatory variables included Our current research covers 226 studies carried out between 1960 and 2008 yielding a total of 1749 valuations (a 50% increase relative to our previous work) and making this the largest data set of its kind to the best of our knowledge This is also the most comprehensive study to date of parameters other than in vehicle time and includes valuations of walk wait headway congested free flow late departure time shift and search time Exploratory analysis of the data set provides interesting insights into methodological trends in travel demand modelling For each valuation over thirty quantitative and categorical variables were recorded and then included in a multivariate regression model to explain variations in the value of time A large number of statistically significant effects were obtained from this meta-analysis some of which are in marked contrast with or not present in our previous work One finding that stands out is that the estimated elasticity of the value of time with respect to GDP per capita is 09 and highly significant a much closer correspondence to the widely used convention of a unit income elasticity over time than we have previously obtained The ratio between walk and wait time and in-vehicle time was found to be substantially lower than the commonly used value of two We also found large and significant differences between the results from studies based on different types of Stated Preference survey presentation Other important effects include variations by mode used mode valued travel purpose attribute type and distance It is envisaged that the results are of direct relevance in the British context as inputs to appraisal or for benchmarking whilst the methodological implications are of broader interest and the results in terms of time equivalents and variations in values of time can be transferred to other contexts (C) 2010 Published by Elsevier Ltd