Modeling Count Outcomes from HIV Risk Reduction Interventions: A Comparison of Competing Statistical Models for Count Responses

Table 2

The estimated parameters, coefficients of the Poisson, NB, ZIP, and ZINB models for all vaginal sex episodes outcome at 12 months.

Parameter	Poisson	Negative binomial	Zero-inflated poisson		Zero-inflated negative binomial
			Poisson	Logit	NB	Logit
Intercept	1.5424*	1.7947*	2.0453	−0.8089	2.0630*	−0.1824
Age	0.0491*	0.0290	0.0392*	−0.0476	0.0252	−0.0781
white	0.1406*	0.1914	0.1077*	−0.0588	0.1863	0.0773
Multiracial	0.4249*	0.3576**	0.3165*	−0.8917	0.2679	−1.0362
Other race	0.0275	−0.0222	0.2054	0.1203	0.0388	0.3403
Hispanic	−0.2908*	−0.1882	−0.1976*	0.5028	−0.1406	0.4350
Poverty	0.2060*	0.2899*	0.1600*	−0.3786	0.2232*	−0.4106
Baseline	0.0163*	0.0181*	0.0147*	−0.0206*	0.0159*	−0.0227
Condition	−0.1542*	−0.2062**	−0.1400*	0.1227	−0.1911*	0.0813
Dispersion		1.3873*			0.8370*
Lagrange multiplier	63120.17*
−2 Log Likelihood	9244.29	3512.26	7344.30		3474.70
Parameter	9	10	18		19
AIC	9262.29	3532.26	7380.00		3512.70
BIC	9299.83	3573.98	7455.10		3591.90
Vuong test			7.6369*		3.2400*

*Indicates that the coefficients are significant at 5%; **the coefficients are significant at 10%.