2

I'm estimating a spatial lag model that tests whether the geographical closeness on stores has an effect on sales.

I have a negative, statistically significant rho coefficient for my spatial weight matrix. I'm a bit confused about the interpretation.

Does this mean that shops that are considered neighbors will negatively impact each other's sales? It seems intuitive but I couldn't find any source to confirm this.

stevensallright
  • 21
  • 2
  • 1
    That could be a reasonable interpretation, but it would be case-specific. Would you expect the stores to be all (or mostly) competitors? But this is probably not the only possibility (and without more details it is impossible to judge how reliable and/or large your correlation is, much less your interpretation). – GeoMatt22 Apr 29 '17 at 03:23
  • The model I'm working on uses only uses stores from the same chain. My goal is to assess the attractiveness of a certain region and check if there might be some cannbalization of sales. – stevensallright Apr 29 '17 at 09:57

0 Answers0