Table I.

The fitted ZIP and hurdle models on the MLS data.

Parameter	ZIP model			Hurdle models
	Estimate	Std. error	p-value	Estimate	Std. error	p-value
The Poisson component
β1 Children of alcoholics	0.0663	0.1360	0.6260	0.0809	0.1056	0.4437
β2 Early onset drinking	−0.0904	0.1619	0.5768	0.2341	0.1060	0.0276
β3 Male	0.1421	0.1289	0.2705	0.1292	0.1029	0.2096
β4 Internalizing	0.0789	0.0630	0.2110	0.0918	0.0501	0.0676
β5 Externalizing	0.1712	0.0673	0.0112	0.1893	0.0578	0.0011
${\sigma }_a^2$	0.3856	0.0880	0.0001	0.1997	0.0503	0.0001
The zero component
γ1 Children of alcoholics	−1.1087	0.7656	0.1481	−0.6016	0.2147	0.0052
γ2 Early onset drinking	−4.4751	1.9402	0.0214	−1.3614	0.2115	0.0001
γ3 Male	0.9166	0.8271	0.2682	0.2317	0.2142	0.2798
γ4 Internalizing	0.2297	0.3417	0.5017	0.1002	0.0983	0.3084
γ5 Externalizing	−1.7627	0.5236	0.0008	−0.9785	0.1131	0.0001
${\sigma }_b^2$	15.4720	10.3614	0.1359	2.2824	0.4173	0.0001
σab	0.6702	0.6152	0.2764	−0.2760	0.1213	0.0232
Goodness of fit	χ2 = 2310.79, df = 3133, p ≈ 1			χ2 = 2685.85, df = 3133, p ≈ 1