Table 1

Interval regression analysis of self assessed health in Great Britain 1979–1995

VariableCoefficientt ratio
Other variables entered in the regression but not shown are (a) year effects; (b) year and income interaction effects; and (c) interaction terms for (country * income * year) effects. The tests of the restrictions at the foot of the table indicate their joint significance. The age variable is (age in years)/100.
Female0.008881.78
Age0.3606113.44
Age2−1.10766−18.07
Age30.7109016.62
Female * Age−0.23135−6.23
Female * Age20.616047.41
Female * Age3−0.44777−7.86
Income0.720.0006916.44
Wales−0.01567−7.37
Wales * Income0.720.000232.48
Scotland−0.00354−2.27
Scotland * Income0.720.000132.15
Constant0.82142202.42
Initial log-likelihood−321329.54
Model log-likelihood−302755.44
Pseudo-R25.78%
Number of observations299,968
Tests of restrictions
Year dummies = 0χ2(16)=61.41p<0.0001
Income coefficient is time invariantχ2(16)=69.33p<0.0001
Wales = Englandχ2(18)=74.27p<0.0001
Scotland = Englandχ2(18)=38.64p=0.0032