Table 2

 Variation of the prevalence of ever regular daily smoking according to age and birth cohort, and the effect of the 1976 Tobacco Control Act (TCA) by sex. For men only the main effects and for women also the interaction term between the TCA and cohort trend (p<0.001) were included. Odd ratios (OR) and 95% confidence intervals (CI) of logit model. Corrected for the bias attributable to declining response rate

VariableMen OR95% CIWomen OR95% CI
*Age—young: if age < 25 then age—young = age − 25; if age ⩾ 25 then age—young = 0 (note that OR gives the relative change during an age interval of one year). †Age—old: if age ⩽ 49 then age—old = 0; if age > 49 then age—old = (age-49)/5 (note that OR gives the relative change during an age interval of five years). ‡Cohort: (birth year −1950)/5 (note that OR gives the relative change during a cohort interval of five years). §Before the effect of the TCA (birth cohorts born 1916–1955).
Age—young*1.09(1.07 to 1.11)1.05(1.03 to 1.07)
Age—old†0.98(0.94 to 1.02)0.93(0.89 to 0.97)
Cohort (trend)‡0.96(0.94 to 0.98)1.30§(1.28 to 1.33)
Deviant cohorts:
    Cohort 1916–251.33(1.18 to 1.49)1.49(1.31 to 1.69)
    Cohort 1931–350.88(0.81 to 0.96)0.97(0.88 to 1.07)
    Cohort 1946–501.09(1.03 to 1.16)1.01(0.96 to 1.07)
TCA 1976:
    before1.001.00
    after0.74(0.68 to 0.81)
Women
    1961–19650.45(0.35 to 0.57)
    1966–19700.34(0.26 to 0.45)
    1971–19750.26(0.19 to 0.36)