Getis ord Gi* z score low values is also high. Is that expected?

Discussion created by deleted-user-xHfmAafpj7ty on Jun 27, 2013
I am working on substance abuse data and the values for drug positives range from 0 to 50 (approx.).

When I run the Getis Ord Gi* Hotspot analysis, I am getting very high z scores for areas with 0 drug positives. Is this an error? Or, is it normal considering that the area(polygon) is surrounded by other areas of high values?

I am using contiguity edges and corners!