This field paper is for discussing the relationship between travel behavior and urban form.1 Part I reviews the related literature and tries to cover the main theories and research in this field. Daily Driving Distance as the variable of interest, is associated with many factors including urban form variables. The discussion will gradually narrow down to the association between driving distance and urban density. Instead of using very high dimension model or synthesized index, Population Density itself may have more potential to explain the driving distance than people ever thought about.

Part II introduces the linear, generalized linear, and non-linear models which is the fundamental methods for the question raised in Part I. Meta-analysis is also a useful method to get more general results based on many studies with the same topic. It can be applied on investigating self-selection and publication bias.

  • Several Ideas

Many disaggregated travel-urban form models without psychological factors are underfitting, the corresponding estimates are biased. When the goal of related studies is to provide evidence for public policy, aggregated analysis may be good enough.

The estimated coefficients are incomparable unless the regression models have the common key specification. Because \(\boldsymbol{\hat\beta}=(\mathbf{X'X})^{-1}\mathbf{X'}\mathbf{y}\) is a function depending on \(\mathbf{X}\). Standardized coefficients or elasticities cannot fix this issues.

Due to the non-linear characteristics of many individual and urban-form factors, identifying the effective ranges of variables is useful for policy making.


Xie, Yihui. 2015. Dynamic Documents with R and Knitr. 2nd ed. Boca Raton, Florida: Chapman; Hall/CRC.
———. 2021. Bookdown: Authoring Books and Technical Documents with r Markdown.

  1. This is a field paper using the bookdown package (Xie 2021), knitr (Xie 2015) and TinyTeX:↩︎