Hi Xander,

Thank you for your response! My one is a little later due to a little holiday...

But it's very interesting. I mainly did it in SPSS because I first exported to Excel to make a pivot table to get the averages from each field. The number of found nears per polygon vary so the only way I think I could do it is this way.

In my research I investige the (potential) relation between the price of Agricultural Lots (the Green Polygons) and the nearbyness and number of inhabitants of Villages (the Orange polygons). I'm also putting in some other variables which theoretecally might cause variance.

I'm also using the "Agricultural Areas" division by the CBS ( CBS StatLine - Gebieden in Nederland 2013 ) to proxy for type of use and quality of the agricultural lots. So for each lot I have a (averaged) weight for the proximity to the nearest maximum of 5 villages within 3 km, with the number of inhabitants in account. By adding a dummy for "Agricultural Area" I can calculate interaction for them ( Weight x AgriculturalDummy ).

You're suggestion on using a sum of instead of an average is very interesting. I think it's reasonable to say that making a sum gives a better weight to nearness to a big city and that's what i'm trying to figure out.

Nice to see that we can help each other further in our calculations! I will head back in a few hours to post some results of both 🙂