Table 3.

Multivariate analysis: DIC values for various models under first- and last-activation schemes, with i.i.d. normal, CAR or GMCAR assumption on spatial frailties

Model	Bivariate	First		Last
	Two independent i.i.d. normal	pD	DIC	pD	DIC
1(a)	${\eta }_{ij}={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{1i};\text{log}({\theta }_{ij})={\theta }_0+{\varphi }_{2i};\mathit{N}~\text{Po}(\theta )$	6.022	3632.070	6.794	3637.110
2	$\begin{array}{l}\text{log}({\theta }_{ij})={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{2i};{\eta }_{ij}={\eta }_0+{\varphi }_{1i};\mathit{N}~\text{Po}(\theta )\hfill \\ \text{Two}\text{independent}\text{CAR}\hfill \end{array}$	5.569	3631.270	3.867	3634.190
1(a)	${\eta }_{ij}={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{1i};\text{log}({\theta }_{ij})={\theta }_0+{\varphi }_{2i};\mathit{N}~\text{Po}(\theta )$	6.311	3632.720	6.880	3637.370
2	$\begin{array}{l}\text{log}({\theta }_{ij})={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{2i};{\eta }_{ij}={\eta }_0+{\varphi }_{1i};\mathit{N}~\text{Po}(\theta )\hfill \\ \text{GMCAR}\hfill \end{array}$	5.965	3632.030	3.954	3634.580
1(a)	${\eta }_{ij}={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{1i};\text{log}({\theta }_{ij})={\theta }_0+{\varphi }_{2i};\mathit{N}~\text{Po}(\theta )$	6.352	3631.580	7.296	3637.050
2	$\text{log}({\theta }_{ij})={\mathbf{x}}_{ij}^{\mathrm{T}}\beta +{\varphi }_{2i};{\eta }_{ij}={\eta }_0+{\varphi }_{1i};\mathit{N}~\text{Po}(\theta )$	6.248	3631.620	1.159	3624.170