1 Reply Latest reply on Mar 11, 2012 12:55 PM by srpang89

    Determining significance of clustering to specific areas?

      Hi guys,

      I'm working with Spatial Statistics for the first time and have a quick question. I have pollution data and have already run a General G to determine if there is a high- or low-cluster pattern and if it is the result of random chance (it was a high-cluster pattern with less than a 1% chance that it is random). I also have data of a disease that could potentially be linked to this pollution. How would I determine if the incidence of this disease around these "hotbeds" of pollution are random? What I'm trying to do, is run General G again, but around specific points (the points where the disease have occurred) if that makes any sense.

      Any help would be GREATLY appreciated!


      Edit: After more researching, should I be performing a Regression Analysis? And should I be using Gi* instead of General G? After reading a bit, General G is best when the measured values are pretty consistent over the entire study area while Gi* is better for larger variations since it looks at neighboring features instead. Even if my measured values are consistent over the study area, could I still use Gi*? Why might I want to use Gi* over General G if my values are consistent?

      Really sorry for all the questions guys. I just feel like all of the "Analyzing Patterns" and "Mapping Clusters" toolsets are all so similar. It's really hard for me to visualize the difference between them without real world examples. It's been hard for me to conceptualize.