Poverty increases going from the northeast of the state toward the southwest; however, prostate cancer incidence rates increase going from the southwest to the northeast. We therefore see an apparent inverse relationship between these two variables.

So, we grossly see apparent patterns in our data. Are these patterns statistically significant, or just random noise? We performed a statistical test (Tango’s Maximum Excess Events Test) to test for any significant clustering of prostate cancer incidence in the state. The Tango test indicated there was statistically significant clustering. Where? We used a local spatial scan statistic (SaTScan) to calculate where local clustering is occurring.