Table 3.

Average length for the bootstrap confidence intervals of the predictiveness curve estimators.

	Method	R(v₁)	R⁻¹(0.1)	R(v₂)	R⁻¹(0.3)
Model 1: μ = 0.5, v₁ = R⁻¹(0.1) = 0.074, v₂ = R⁻¹(0.3) = 0.886
n_D = n_D̄ = 250	PROC¹	0.051	0.26	0.058	0.18
	PROC²	0.065	0.42	0.060	0.26
n_D = n_D̄ = 500	PROC¹	0.038	0.19	0.045	0.13
	PROC²	0.047	0.27	0.046	0.18
Model 2: μ = 0.8, v₁ = R⁻¹(0.1) = 0.232, v₂ = R⁻¹(0.3) = 0.808
n_D = n_D̄ = 250	PROC¹	0.047	0.21	0.065	0.11
	PROC²	0.057	0.29	0.070	0.17
n_D = n_D̄ = 500	PROC¹	0.035	0.15	0.051	0.081
	PROC²	0.041	0.19	0.053	0.11
Model 3: μ = 1.2, v₁ = R⁻¹(0.1) = 0.396, v₂ = R⁻¹(0.3) = 0.770
n_D = n_D̄ = 250	PROC¹	0.04	0.20	0.082	0.073
	PROC²	0.045	0.26	0.097	0.14
n_D = n_D̄ = 500	PROC¹	0.029	0.14	0.064	0.058
	PROC²	0.032	0.16	0.074	0.086

partial binormal ROC modeling with t ∈ (0.01, 0.99)

partial binormal ROC modeling with t ∈ (0.50, 0.99) for estimating R⁻¹(0.1) and R(v₁) for v₁ = R⁻¹(0.1), and t ∈ (0.01, 0.50) for estimating R⁻¹(0.3) and R(v₂) for v₂ = R⁻¹(0.3)