Table 3.

Performance of the MCEM algorithms: mean estimates, standard deviations (SD) and 95% CI coverages are based on 500 simulated 30-day epidemics among 100 households each of size 5, with b = 0.01, p = 0.05, α = 0.2, φ = 0.3 and θ = 0.5. θ is assumed known. Burn-in and final numbers of importance samples are K₀ = 200 and K = 1000.

Assumed value of θ	Parameter	True Value	ξ = 0.25			ξ = 0.5			ξ = 0.75
			Mean	x SD	95% CI Coverage	Mean	x SD	95% CI Coverage	Mean	x SD	95% CI Coverage
Traditional MCEM (J = 2)
0.5	b	0.01	0.010	0.0015	94.8%	0.010	0.0015	94.4%	0.010	0.0016	95.6%
	p	0.05	0.050	0.0098	94.6%	0.050	0.010	95.2%	0.050	0.011	95.6%
	φ	0.70	0.70	0.041	93.4%	0.70	0.041	93.4%	0.70	0.041	93.4%
	α	0.20	0.20	0.020	94.0%	0.20	0.024	94.8%	0.20	0.034	95.8%
	OR	0.60	0.61	0.11	94.8%	0.61	0.12	95.4%	0.61	0.12	95.2%
Hybrid EM-MCEM (J = 1000)
0.5	b	0.01	0.010	0.0015	94.8%	0.010	0.0015	94.4%	0.010	0.0016	95.6%
	p	0.05	0.050	0.0098	94.6%	0.050	0.010	95.2%	0.050	0.011	96.0%
	φ	0.70	0.70	0.041	93.4%	0.70	0.041	93.4%	0.70	0.041	93.4%
	α	0.20	0.20	0.020	93.8%	0.20	0.024	95.2%	0.20	0.033	95.4%
	OR	0.60	0.61	0.11	94.8%	0.61	0.12	95.4%	0.61	0.12	95.2%
0	b	0.01	0.011	0.0015	91.4%	0.011	0.0016	92.2%	0.011	0.0016	92.6%
	p	0.05	0.055	0.011	92.6%	0.055	0.011	93.2%	0.055	0.012	92.8%
	φ	0.70	0.70	0.041	93.4%	0.70	0.041	93.4%	0.70	0.041	93.4%
	α	0.20	0.20	0.020	94.0%	0.20	0.024	95.8%	0.20	0.033	95.6%
	OR	0.60	0.61	0.11	95.0%	0.61	0.12	95.4%	0.61	0.12	95.0%
1	b	0.01	0.0099	0.0015	95.2%	0.0099	0.0015	95.6%	0.010	0.0016	95.4%
	p	0.05	0.044	0.0085	91.0%	0.044	0.0087	90.2%	0.044	0.0091	92.4%
	φ	0.70	0.70	0.041	93.4%	0.30	0.041	93.4%	0.70	0.041	93.4%
	α	0.20	0.20	0.020	94.0%	0.20	0.024	95.2%	0.20	0.033	95.4%
	OR	0.60	0.61	0.11	95.0%	0.61	0.12	95.4%	0.61	0.12	95.2%