Skip to main content
. 2007 Jun;61(6):499–505. doi: 10.1136/jech.2006.049940

Table 3 Regression‐based relative risks for educational and social class differences in mortality, adjusted for 5‐year age groups and 5‐year periods, for men.

1971–1980 1981–1990 1991–2000 1991–2000 with 1971–1980 population distribution*
Total effect† Independent effect‡ % Independent§ Total effect† Independent effect‡ % Independent§ Total effect† Independent effect‡ % Independent§ Total effect† Independent effect‡ % Independent§
Age 30–39 years
Education 1.66 1.43 65.4 1.67 1.37 54.6 2.04 1.68 65.7 2.21 1.81 66.8
Social class 1.55 1.27 48.0 1.71 1.47 66.6 1.87 1.52 60.0 1.96 1.42 44.1
Age 40–49 years
Education 1.52 1.30 58.3 1.50 1.21 41.2 1.67 1.34 50.7 1.79 1.39 49.0
Social class 1.45 1.24 53.1 1.56 1.41 72.9 1.71 1.49 69.2 1.74 1.45 60.1
Age 50–59 years
Education 1.39 1.27 68.0 1.41 1.20 47.8 1.44 1.16 36.0 1.49 1.17 34.7
Social class 1.30 1.14 45.8 1.39 1.26 66.1 1.53 1.41 77.8 1.53 1.40 75.6

*Hypothetical values based on calculations that keep the education and social class population distributions, and the correlations between the two constant at the 1971–80 level (see Methods).

†Total effect  =  total effect of education or social class on mortality.

‡Independent effect  =  effect of education independent of social class or effect of social class independent of education.

§% independent  =  independent effect as a proportion of the total effect (see Methods).