(ENG) In this article we suppose that the integration of accessibility and neighborhood variables in hedonic housing models doesn't allow to take into account the spatial effects between the housing prices. Using a sample of 1520 transactions of apartments in the urban area of Dijon, we focus on two types of location variables : the distance to the CBD and the distance to several Disadvantaged Districts located in peripheral areas. We detect the presence of a spatial error autocorrelation in the hedonic model indicating that the valuation of the price of an apartment is locally influenced by the prices of the neighboring apartments. Then, we estimate a spatial error model which shows that the local effect is positive and significant and tha...