Table 5.

Fitted log-linear and POMs for numiadl, AHEAD data. Log-linear models are estimated with SPGLM and quasi-Poisson regression

	Log-linear mean
	SPGLM			Quasi-Poisson		POM
		()	Z		()		()	Z
Intercept	– 3.61	0.337	– 10.69	– 3.62	0.337	—	—	—
age	0.05	0.004	12.74	0.05	0.004	0.07	0.005	12.71
sex:female	0.12	0.052	3.04	0.16	0.052	0.26	0.064	4.01
iwr	– 0.21	0.014	– 14.72	– 0.21	0.014	– 0.26	0.018	– 14.42
netwc:1–24k	– 0.26	0.078	– 3.28	– 0.26	0.077	– 0.45	0.113	– 4.01
netwc:25–74k	– 0.45	0.080	– 5.62	– 0.46	0.079	– 0.65	0.111	– 5.81
netwc:75–199k	– 0.69	0.081	– 8.50	– 0.69	0.081	– 0.93	0.110	– 8.46
netwc:200k–up	– 0.76	0.090	– 8.48	– 0.76	0.090	– 0.92	0.116	– 7.90
Log-likelihood	– 4951.2					– 4951.5
Scale					1.48