In 1976, the Institute of Transportation Engineers (ITE) compiled their first Handbook of guidelines and methods for evaluating development-level transportation impacts, specifically vehicular impacts (Institute of Transportation Engineers 1976). Decades later, these methods— essentially the same as when they were originally conceived—are used ubiquitously across the US and Canada. Only recently, with the guidelines in its third edition of the ITE’s Trip Generation Handbook (Institute of Transportation Engineers 2014) new data and approaches been adopted—despite substantial evidence that questions the accuracy of older data (Clifton, Currans, and Muhs 2012; Shafizadeh et al. 2012; Weinberger et al. 2015), automobile bias (Clifton et al. 2012; Millard-Ball 2015; Manville 2017), and lack of sensitivity to urban contexts (Currans and Clifton 2015; Ewing et al. 2011; Schneider, Shafizadeh, and Handy 2015; Weinberger et al. 2015). 

This dissertation contributes to this literature by focusing on the data, methods, and assumptions so commonly included in development- or sitelevel evaluation of transportation impacts. These methods are omnipresent in development-level review—used in transportation impact analyses or studies (TIAs/TISs) of vehicular or mode-based impacts, vehicle miles traveled (VMT) and estimates of emissions, scaling or scoping development size, and evaluating transportation system development, impact or utility fees or charges. However, few have evaluated the underlying characteristics of these foundational data—with few exceptions (Shoup 2003)—this manuscript takes aim at understanding inherent issues in the collection and application of ITE’s data and methods in various urban contexts. 

This manuscript includes a compiled dissertation, four papers written consecutively. The first, evaluates state-of-the-art methods in Chapter 2— identifying gaps in the literature. Two such gaps are explored in Chapter 3 and Chapter 4. In Chapter 3, a larger implicit assumption present in ITE’s methods—that the existing land-use taxonomy is an optimal and accurate way to describe land use and segment data. Results indicate a simplified taxonomy would provide substantial reductions in cost corresponding with a minor loss in the model’s explanation of variance. Following, Chapter 4 explores a common assumption that requires ITE’s vehicle trips be converted into person trips and applied across contexts. The results point to the need to consider demographics in site-level transportation impact analysis, particularly to estimate overall demand (person trips, transaction activity) at retail and service development. 

In Chapter 5, the findings from this research and previous studies are extrapolated to evaluate and quantify the potential bias when temporal, special, and social contexts are ignored. The results indicate the compounding overestimation of automobile demand may inflate estimation by more than 100% in contexts where ITE should be applicable (suburban areas with moderate incomes). In the conclusions (Chapter 6), the implications of this work are explored, followed by recommendations for practice and a discussion of the limitations of this research and future work.