Skip to main content

Advertisement

Table 5 Number of counties where the difference between cancer mortality rates is declared significant (two-tailed test). Results are reported for the two test statistics that performed best in simulation studies. Significance levels are either unadjusted or reduced using the Bonferroni correction, the Holm's procedure or the False Discovery Rate (FDR) approach. Agreement refers to the number of counties declared significant by both tests.

From: A comparative analysis of aspatial statistics for detecting racial disparities in cancer mortality rates

Test statistic α = 0.01 α = 0.05
  Unadj. Bonf. Holm FDR Unadj. Bonf. Holm FDR
Prostate         
DispI 385 176 184 362 462 213 222 430
DispVI 372 167 170 350 452 200 209 422
Agreement 372 167 170 350 452 200 209 422
Lung         
DispI 70 13 13 26 136 19 19 48
DispVI 66 13 13 25 134 19 19 48
Agreement 66 13 13 25 134 19 19 48