Skip to main content
. 2014 Jun 23;25(1):44–49. doi: 10.1093/eurpub/cku078

Table 4.

Relationship between number of risk behaviours and measures of decreasing SES (imputed sample, n = 6406)

SES measure n Number of risk behaviours
Ordinal logistic regression models
0–1 2–3 4–6 7–13 Unadjusted OR [95% CI] Mutually adjusted OR [95% CI]
Parental social class
    Professional 1089 28.8% 42.2% 23.9% 5.1% 1.0 ref 1.0 ref
    Managerial 2883 21.9% 40.3% 29.2% 8.7% 1.22 [1.15, 1.29], P < 0.001 1.06 [1.02, 1.11], P = 0.008
    Skilled non-manual 1537 20.8% 40.5% 31.0% 7.8%
    Skilled manual and lower 897 16.4% 39.7% 31.5% 12.5%
Maternal educational attainment
    Degree 1153 27.4% 40.3% 25.9% 6.4% 1.0 ref 1.0 ref
    A-level 1730 23.6% 40.0% 28.6% 7.9% 1.15 [1.09, 1.21], P < 0.001 1.04 [0.98, 1.11], P = 0.184
    O-level 2178 19.8% 41.1% 30.0% 9.1%
    <O-level 1345 19.0% 40.6% 30.7% 9.7%
Quintiles of household equivalized income
    High 1537 24.9% 41.2% 27.4% 6.6% 1.0 ref 1.0 ref
    Middle high 1409 24.8% 39.8% 28.3% 7.1% 1.12 [1.08, 1.16], P < 0.001 1.15 [1.07, 1.23], P < 0.001
    Middle 1281 22.3% 41.1% 27.5% 9.0%
    Middle low 1153 18.6% 40.4% 31.4% 9.7%
    Low 1025 17.4% 40.2% 31.7% 10.7%

Ordinal regression models under the proportional odds assumption with a linear relationship for each SES predictor variable. P-values for contingency tables (not shown) derived from Chi-square statistics were all <0.001. P-values shown are derived from Wald tests. Percentages shown are row percents.

For the ‘mutually adjusted’ estimates, the effect of each SES measure has been adjusted for the other two SES measures.