Comparison of tests for spatial heterogeneity on data with global clustering patterns and outliers

Table 2 Power of selected statistics for detecting data with local and global cluster types.

		Global clustering methods						Local cluster detection methods
		Tango_PDM	Tango_ADJ	Moran's I ADJ	I *_{pop _}PD ( λ = 5%)	I *_{pop _}PD ( λ = 50%)	I *_{pop _}ADJ	Local Moran's I	SaTScan^a
Local clusters relative risk = 4 sample size = 30,000	Two adjacent neighbors (Manassas City VA)	0.107	0.127	0.024	.079	0.056	0.060	+	0.973
	Six adjacent neighbors (Jackson County KS)	0.089	0.124	0.027	.087	0.042	0.047	+	0.962
	Ten adjacent neighbors (Fulton County GA)	0.081	0.125	0.030	.079	0.043	0.055	+	0.979
Global clustering relative risk = 1.5 sample size = 5.000	Linear	0.997	0.285	0.232	.884	0.996	0.119	+	0.737
	Exp	0.998	0.291	0.235	.886	0.997	0.129	+	0.781

^a for global clustering evaluation, we used 50% of total population as the maximum search window. With the outlier detection, we use 5% of the total population as the maximum search window in SaTScan-E.
+ Power calculations are not available for Local Moran's I due to the fact that a statistics is produced for each geographic unit in the study. It is difficult to define power in the presence of multiple comparisons.

ISSN: 1476-072X