Table 3.

For Design 4, θ ₁ = 0.5, θ ₂ = 0, and θ ₃ = 0) standardized bias and standardized $\sqrt{\text{MSE}}$ of the MLEs where the prevalence rates of three subgroups are 1/3. In addition, Proportion (Prop.) stands for how often the corresponding circumstance occurs

	Stop at Stage 1			Stop at Stage 2			Overall
	Bias/SE	$\sqrt{\text{MSE}}/\text{SE}$	Prop.(%)	Bias/SE	$\sqrt{\text{MSE}}/\text{SE}$	Prop.(%)	Bias/SE	$\sqrt{\text{MSE}}/\text{SE}$	Prop.(%)
${\widehat{\theta }}_f$ (Select none)	0.67715	1.13140		‐0.26340	0.96585		0.05614	1.02209
${\widehat{\theta }}_f$ (Select F )	2.02110	2.14318	1.54	0.36337	0.92461	5.98	0.70283	1.17414	7.52
${\widehat{\theta }}_f$ (Select F + reject H 0F)	2.10568	2.14995	1.51	0.85268	1.01727	3.89	1.20283	1.33379	5.39
${\widehat{\theta }}_1$ (Select none)	1.03255	1.22555		‐0.29968	0.97840		0.15293	1.06237
${\widehat{\theta }}_1$ (Select S 1 )	1.12932	1.27694	29.13	‐0.20496	0.94416	51.17	0.27906	1.06487	80.29
${\widehat{\theta }}_1$ (Select S 1 + reject H 01)	1.14828	1.26420	28.99	‐0.20175	0.93853	51.11	0.28684	1.05639	80.11
${\widehat{\theta }}_{1+2}$ (Select none)	0.81892	1.16958		‐0.26809	0.96219		0.10120	1.03265
${\widehat{\theta }}_{1+2}$ (Select S 1+2)	1.78983	1.88884	3.31	0.17732	0.89687	8.88	0.61488	1.16604	12.18
${\widehat{\theta }}_{1+2}$ (Select S 1+2 + reject H 0,1+2)	1.82834	1.88619	3.27	0.41744	0.81323	7.57	0.84338	1.13715	10.85
Family‐wise select	1.23403	1.37576	33.97	‐0.10207	0.93603	66.03	0.35185	1.08542
Family‐wise select + reject	1.25693	1.36402	33.77	‐0.06135	0.92826	62.57	0.40076	1.08101

Abbreviations: MSE, mean squared error.