Skip to main content

Table 9 Five measures on the income-health relationship from fitting a lognormal distribution to the transformed health variable using GAMLSS regression

From: A distributional regression approach to income-related inequality of health in Australia

 

$20,000

$80,000

Absolute difference

Relative difference

a) Smokers

E(h)

0.759 [0.733; 0.773]

0.770 [0.742; 0.784]

0.011 [0.001; 0.031]

1.41% [ 0.08%; 3.99%]

σ

0.194 [0.148; 0.266]

0.188 [0.143; 0.263]

0.006 [0.001; 0.059]

3.19% [ 0.34%; 24.25%]

P avg

0.465 [0.452; 0.489]

0.436 [0.426; 0.458]

0.028 [0.012; 0.046]

6.06% [ 2.57%; 9.66%]

P fair

0.254 [0.223; 0.290]

0.233 [0.200; 0.274]

0.020 [0.001; 0.056]

8.03% [ 0.33%; 20.26%]

P poor

0.142 [0.107; 0.183]

0.129 [0.093; 0.173]

0.013 [0.001; 0.049]

9.29% [ 0.47%; 29.94%]

b) Non-smokers

E(h)

0.774 [0.756; 0.790]

0.784 [0.764; 0.801]

0.010 [0.001; 0.028]

1.30% [ 0.07%; 3.47%]

σ

0.182 [0.133; 0.236]

0.176 [0.129; 0.233]

0.006 [0.000; 0.051]

3.20% [ 0.30%; 23.91%]

P avg

0.429 [0.408; 0.447]

0.401 [0.382; 0.419]

0.027 [0.009; 0.050]

6.39% [ 2.03%; 11.47%]

P fair

0.226 [0.184; 0.256]

0.207 [0.162; 0.242]

0.019 [0.001; 0.055]

8.34% [ 0.25%; 23.44%]

P poor

0.123 [0.082; 0.155]

0.111 [0.070; 0.146]

0.012 [0.001; 0.045]

9.57% [ 0.52%; 33.59%]

  1. Note. 95th percentile bootstrap confidence intervals denoted in brackets