Table 2.
Global tests for spatial randomness applied to brain cancer mortality in the United States, 1986 to 1995
|
Tango’s EET |
Cuzick and Edwards k-NN |
|||
|---|---|---|---|---|
| λ (km)* | P-value | K† | P-value | |
| All Adults (20 and up) | 1,000 | 0.0001 | 4,000 | 0.0001 |
| 2,000 | 0.0001 | 5,000 | 0.0001 | |
| 5,000 | 0.0001 | 8,000 | 0.0001 | |
| 10,000 | 0.0001 | 10,000 | 0.0001 | |
| Men (20 and up) | 1,000 | 0.0001 | 1,800 | 0.0001 |
| 2,000 | 0.0001 | 2,000 | 0.0001 | |
| 5,000 | 0.0001 | 3,000 | 0.0001 | |
| 10,000 | 0.0001 | 4,000 | 0.0001 | |
| Women (20 and up) | 1,000 | 0.0001 | 1,500 | 0.0001 |
| 2,000 | 0.0001 | 2,000 | 0.0001 | |
| 5,000 | 0.0001 | 2,500 | 0.0001 | |
| 10,000 | 0.0001 | 3,000 | 0.0001 | |
| Children (under 20) | 1,000 | 0.005 | 200 | 0.04 |
| 2,000 | 0.06 | 400 | 0.08 | |
| 5,000 | 0.21 | 500 | 0.13 | |
| 10,000 | 0.29 | 600 | 0.17 | |
Abbreviations: EET, Excess Events test; k-NN, k-Nearest Neighbor (test).
*
The λ used for Tango’s method (Tango, 1995) is defined in Eq. 1.
†
The parameter k defines the size of the circles used by the Cuzick-Edwards method (Cuzick and Edwards, 1990) to define neighbors, measured in terms of the expected number of cases under the null hypothesis.