. 2012 Jan 21;2(1):e000512. doi: 10.1136/bmjopen-2011-000512

Table 3.

Log-binomial regression models. Pseudo R² and differences in pseudo R² between models with and without occupation among women

Age group	Model variables
	Age	Age and occupation		Age, occupation and occupational noise exposure		Age, occupation and all risk factors^†		Age, occupation, all risk factors^†, income and education		Age, occupation, all risk factors^†, income, education and hearing loss
	$R_{l}^{2}$ (%)	$R_{l}^{2}$ (%)	$Δ R_{l}^{2}$ (%)	$R_{l}^{2}$ (%)	$Δ R_{l}^{2}$ (%)	$R_{l}^{2}$ (%)	$Δ R_{l}^{2}$ (%)	$R_{l}^{2}$ (%)	$Δ R_{l}^{2}$ (%)	$R_{l}^{2}$ (%)	$Δ R_{l}^{2}$ (%)
All	2.3	2.8	0.59**	3.4	0.62**	5.1	0.59**	5.4	0.38	7.8	0.45**
20–44 years	0.0	1.3	1.31*	2.7	1.38**	4.6	1.20	5.4	1.01	7.4	0.97
45–64 years	0.3	1.4	1.15*	1.8	1.17*	3.9	1.19*	4.1	1.02	7.3	1.08**
>64 years	0.5	1.4	0.91*	2.2	0.99	3.7	1.04	4.1	0.77	7.0	0.85**

$R_{l}^{2}$ , pseudo R² (McFadden, 1979); $Δ R_{l}^{2}$ , partial pseudo R² is the difference in $R_{l}^{2}$ values between a model with and without occupation.

*p<0.05; **p<0.01, likelihood ratio test on 54 df.

^†

Self-reported occupational noise, leisure noise, recurrent ear infections, head injuries and smoking.