I am working on a project where we are measuring the degree of clustering within a city for 1,300 incidents. I was using the "Average Nearest Neighbor" tool to calculate z- and p-scores, which worked fine. Now, I wan't to compare these values to the actual population of the city, to see if the population or my incidents are clustered at a higher degree (similar as described in the ESRI Turorial on Dengue Fever http://www.arcgis.com/home/item.html?id=020e1d9e9d2d44f9ae7da8663cf3bac5 - page 7 of the PDF).
The problem is, that I have the population not per incident, which would make no sense, but on a fishnet based on 100 x 100 meters. I can also run "Averagy Nearest Neighbor" on my fishnet polygons, but I have a feeling, that this is no valid comparison and that it would make more sense to use a spatial join to count the incidents per fishnet cell and to perform another statistical test - right?
If yes, which other tool(s) than "Average Nearest Neighbor" would you recommend to calcluate the clustering of my fishnet count values (number of incidents vs. number of citizens)?
Thanks a lot for your help,