. 2018 Aug 21;15(9):1792. doi: 10.3390/ijerph15091792

Table 4.

The comparison of model fitting for the Poisson Hurdle and negative binomial (NB) Hurdle models.

Variable	Model Parameter	Poisson Hurdle with RE		NB Hurdle with RE
Variable	Model Parameter	Estimate	SE	Estimate	SE
	${\hat{α}}_{0}$	31.273 **	11.143	25.016 *	11.099
Building Year	${\hat{α}}_{1}$	−0.017 **	0.006	−0.014 *	0.006
Town Taxable Value	${\hat{α}}_{2}$	−0.731 *	0.305	−0.787 *	0.304
Soil Lead	${\hat{α}}_{3}$	0.588 ***	0.162	0.621 ***	0.162
	${\hat{β}}_{0}$	23.746 *	9.150	24.003	13.959
Building Year	${\hat{β}}_{1}$	−0.013 **	0.005	−0.013	0.007
Town Taxable Value	${\hat{β}}_{2}$	−0.562 *	0.247	−0.927 **	0.335
Soil Lead	${\hat{β}}_{3}$	0.182	0.112	0.288	0.179
	${\hat{σ}}_{1}^{2}$	0.693 ***	0.120	0.687 ***	0.119
	${\hat{σ}}_{2}^{2}$	0.439 ***	0.100	0.346 *	0.147
	κ			1.259 *	0.484

Note: ${\hat{α}}_{0}$ ~ ${\hat{α}}_{3}$ are the model coefficients for the logistic regression in the Hurdle models; ${\hat{β}}_{0}$ ~ ${\hat{β}}_{3}$ are the model coefficients for the truncated Poisson or NB Hurdle models; The symbols *, **, and *** indicate statistical significance at α = 0.05, 0.01, and 0.001 levels, respectively.