ABSTRACT
This chapter examines spatial dependence in cross-sectional data or panel data and explore their causes and the effects they have on the estimation of econometric models. The spatial effects present in cross-sectional data or in panel data can be classified into two general types: Spatial dependence and Spatial heterogeneity. Spatial dependency between each household's trip production can be interpreted as a diffusion effect from similar mobility patterns between nearby households. The chapter describes the importance of considering the possible presence of spatial dependence effects in the data used for estimating different sub-models within a land use–transport interaction (LUTI) model. This spatial dependence may be commonly present in the cross-sectional data and the panel data, thereby making the parameters of the estimated regression or discrete choice models biased and/or inefficient. In the case of the spatial autoregressive model, the prediction should consider the relationship of spatial dependence between the trips generated from neighbouring observations.
