Figure 1.

The total sample sizes of the full population F ( $n_f^{(1)}$) across prevalence rates of S ₁ (λ₁) for two different definitions of power. The design is a single‐stage design with two subpopulations where the treatment effects θ ₁ and θ ₂ for S ₁ and S ₂ are 0.5 and 0, respectively. The type‐I error and power are specified at 0.025 and 80% [Colour figure can be viewed at http://wileyonlinelibrary.com]