The structural assumptions embedded in a particular form of a spatial connectivity or dependence matrix are often not explicitly treated in the statistical analysis of mixed regressive-spatial autoregressive models. This paper sees the evaluation of different structures as an application of tests on non-nested hypotheses. Several recently developed statistics are outlined and their appropriateness for spatial analysis discussed. The tests are illustrated empirically in simple models that incorporate spatial spill-over effects in the explanation of urban housing value.
ASJC Scopus subject areas
- Geography, Planning and Development
- Environmental Science (miscellaneous)