Trend-Breaking Influences of Built Form on Travel in UK Cities Evidence from New Quantifications of Within- and Between-Built-Form Variations

Using a consistent time series of 11 years (1,137,259 sample trips, 2002 to 2012) and all main variables from the UK National Travel Survey, this paper develops a novel structural equation model (SEM) with random intercepts to gain new insights into the influences of built-form characteristics on travel behavior. With simultaneous modeling of car ownership as an endogenous variable in determining influences on travel outcomes and allowing SEM intercepts to vary across built-form categories, one controls for car ownership endogeneity while measuring within- and between-built-form variations through incorporating random intercepts in the model. The new models unambiguously decompose the self-selection and spatial sorting of household effects from the influence of the built-form characteristics. This research further investigates trend-breaking influences by generating pre-2007 and post-2007 model outputs with a multi-time-period structure. The new quantifications are built on more-robust statistical theories than hitherto and, thus, imply significant implications on how to interpret the effect of recent transformations in UK inner cities. The method can also be used to produce a regular and timely update on any shifts in the influences on travel and provide feedback for land use planning and integrated built-form and transport interventions