Skip to main content
. 2014 Oct 29;13(Suppl 7):1–10. doi: 10.4137/CIN.S16352

Figure 1.

Figure 1

Average power at nominal α = 0.05. Sample size (n1, n2, n3) and effect of confounding β are indicated in the header of each plot. Power curves for methods that did not control Type I error (ie, empirical Type I error > 0.055) are not shown.

Notes: ○: Gold-std, ∆: FM-PS, +: Reg-PS, ×: two-sample, and ⋄: paired only.