Table 2.

Estimated power, sensitivity and positive predictive value (PPV) when the normal scan statistic is used to detect different types of clusters, as described in the text.

Cluster number	Cluster size (%)	Sample size	Cluster mean	Cluster STD	Power (%)	Sensitivity	PPV
Different Cluster Locations
1	10	600	-10%	600	66	0.70	0.74
2	10	600	-10%	600	54	0.56	0.65
3	10	600	-10%	600	64	0.72	0.82
4	10	600	-10%	600	61	0.61	0.64
Different Cluster Size
1	5	600	-10%	600	34	0.52	0.48
1	10	600	-10%	600	66	0.70	0.74
1	20	600	-10%	600	96	0.87	0.89
Different Sample Size
1	10	300	-10%	600	34	0.55	0.56
1	10	600	-10%	600	66	0.70	0.74
1	10	900	-10%	600	89	0.83	0.82
1	10	1200	-10%	600	98	0.85	0.85
Different Mean Weight Reduction
1	10	600	-5%	600	14	0.30	0.38
1	10	600	-8%	600	42	0.56	0.61
1	10	600	-10%	600	66	0.70	0.74
1	10	600	-13%	600	92	0.84	0.84
1	10	600	-15%	600	98	0.88	0.88
1	10	600	-20%	600	100	0.93	0.93
Different Standard Deviation
1	10	600	-10%	300	100	0.93	0.93
1	10	600	-10%	600	66	0.70	0.74
1	10	600	-10%	900	28	0.45	0.51