T'étais la seule personne qui me voyais tel que j'étaisinternally homogeneous. Montpellier MSE. But this problem is irrelevant to certain types of spatial analysis : ecological analysis sensu stricto and certain examples of factorial ecologies.
Some clearly are of type B in Dogan and Rokkan's categorization, and so, according to the arguments presented here, are immune from the spatial autocorrelation problem. Thus, as Dogan and Rokkan put it in their introduction to Part 2 of the book, surveys should be ecologically controlled -so as to provide data for type D analyses in particular -rather than the atomistic collection of individual data characteristic of most national surveys.
We are not pretending that the outcome of evolution was fully determined or predictable, but we want to argue against the supposition that all things are possible. Our culture is built on the assumption that we are the superior life form on Earth. This allows them to classify data into four types.
Site de la documentation.