Table 1.

Mean estimates of p and λ for models I and II over 100 simulations for N = 1000 observations simulated from “a mixture of 2 bivariate normal distributions with constant p.” μ₀ = (10,10), μ₁ = (11, 11), Σ₀ = [0.25 0.1; 0.1 0.25], and Σ₁ = [0.25 0.05; 0.05 0.50]. Empirical standard errors are given in parenthesis

Model	λ1 = 1	λ2 = l	p = 0.25	Coverage† for p	Coverage‡ for p	Log-likelihood
I, K = 0	1.00 (0.03)	1.00 (0.02)	0.25 (0.06)	0.93	0.94	−2010.71 (32.9)
I, K= 1	1.00 (0.02)	1.02 (0.01)	0.25 (0.06)	0.91	0.93	−2009.69 (32.7)
I, K= 2	1.00 (0.02)	1.00 (0.02)	0.27 (0.11)	0.72	0.71	−2006.68 (32.8)
II, K= 1			0.25 (0.07)	0.93	0.90	−2009.50 (32.6)
II, K= 2			0.10 (0.14)	0.71	0.75	−2007.83 (30.9)
Model	λ1 = 1	λ2 = l	p = 0.25	Coverage for p	Coverage for p	Log-likelihood
I, K = 0	1.00 (0.02)	1.00 (0.02)	0.06 (0.06)	0.78	0.89	−1865.70 (31.0)
I, K = 1	1.00 (0.02)	1.02 (0.02)	0.06 (0.05)	0.85	0.84	−1864.69 (30.9)
I, K = 2	1.00 (0.02)	1.00 (0.02)	0.08 (0.10)	0.64	0.68	−1860.14 (31.2)
II, K = 1			0.08 (0.10)	0.54	0.50	−1860.14 (31.2)
II, K = 2			0.08 (0.17)	NA	0.35	−1862.83 (31.3)

^†Coverage of confidence intervals based on the asymptotic normality of p^.

^‡Coverage of likelihood-ratio test-based confidence intervals.