Figure 4.

(For Design 1, θ ₁ = 0.5 and θ ₂ = 0.5) the standardized bias and standardized $\sqrt{\text{MSE}}$ of maximum likelihood estimators ${\widehat{\theta }}_f$, ${\widehat{\theta }}_1$ and the simulation proportions for different circumstances against the prevalence of subpopulation 1, ie, λ₁. FW, family‐wise; MSE, mean squared error [Colour figure can be viewed at http://wileyonlinelibrary.com]