Table III.

Mean squared error with varied proportions of missing data (r = 0.25, N = 200).

Parameter	Poisson regression model			ZIP model			Hurdle model
	p = 0.20	p = 0.30	p = 0.40	p = 0.20	p = 0.30	p = 0.40	p = 0.20	p = 0.30	p = 0.40
The Poisson component
β1	0.0601	0.0675	0.0635	0.0038	0.0042	0.0058	0.0036	0.0039	0.0043
β2	0.2720	0.2848	0.3036	0.0047	0.0057	0.0085	0.0044	0.0044	0.0069
β3	0.0221	0.0252	0.0269	0.0028	0.0043	0.0075	0.0027	0.0033	0.0052
β4	0.0104	0.0118	0.0114	0.0035	0.0050	0.0066	0.0033	0.0038	0.0056
β5	0.1403	0.1529	0.1512	0.0038	0.0082	0.0095	0.0033	0.0050	0.0050
g(t)*	114.49	147.67	160.37	21.30	22.56	21.83	25.41	21.53	21.45
${\sigma }_a^2$	0.1822	0.1709	0.1365	0.7517	0.7663	0.7422	0.7330	0.7708	0.7795
The zero component
γ1	N/A	N/A	N/A	0.0785	0.1446	0.2256	0.0130	0.0145	0.0199
γ2	N/A	N/A	N/A	0.1759	0.3632	0.5452	0.0186	0.0245	0.0280
γ3	N/A	N/A	N/A	0.0776	0.1361	0.1664	0.0109	0.0143	0.0183
γ4	N/A	N/A	N/A	0.0538	0.0721	0.1156	0.0110	0.0136	0.0192
γ5	N/A	N/A	N/A	0.1133	0.2396	0.3388	0.0157	0.0187	0.0221
h(t)*	N/A	N/A	N/A	15.16	18.42	23.57	12.18	13.12	13.57
${\sigma }_b^2$	N/A	N/A	N/A	1.6315	1.7310	1.9227	1.0753	1.0362	1.1108
σab	N/A	N/A	N/A	0.2874	0.3449	0.5669	0.0350	0.0534	0.0649

The average of the integrated mean squared error was calculated for the spline function.