Background Evidence of associations between socioeconomic position (SEP) and alcohol use in adolescence and young adulthood is inconsistent, possibly because causal mechanisms associated with SEP pull in different directions. For example, more affluent young people may have greater resources for obtaining alcohol, and may drink for pleasure, whilst socioeconomic adversity may increase coping-motivated drinking. Our previous research on adolescents showed disadvantaged SEP was associated with higher levels of drinking in combination with early smoking, but lower levels of drinking without smoking. We investigate whether such opposing processes contingent on early smoking are present in multiple contexts and across the transition into adulthood.

Methods Data were taken from the British 1970 birth cohort study (BCS70; N = 11,649), 1958 National Child Development Study (NCDS58; N = 15,242), and the youth cohort of the Scottish Twenty-07 Study (T07; N = 1,515). T07 represents experiences in the specific context of Glasgow at the same historical time as BCS70, whilst NCDS58 respondents grew up 12 years earlier. Measures were obtained when respondents were aged approximately 16 years and in young adulthood, when respondents were aged approximately 22–26 years across the three studies. Path analyses tested direct associations between parental social class at age 16 and both weekly drinking at age 16 and heavy drinking in young adulthood (>14/21 units last week for women/men), adjusting for an indirect path via early smoking (defined as daily smoking at age 16).

Results In all three studies, a manual compared to a non-manual parental social class was associated with a higher likelihood of daily smoking at age 16, and daily smoking at 16 was associated with a higher likelihood of weekly drinking at 16. Adjusting for the indirect relationship via smoking, a manual social class was directly associated with a lower likelihood of weekly drinking at 16 (ORs ranged between 0.51-0.70; p < 0.05). Daily smoking at 16 also exhibited positive associations with young adult drinking. Adjusting for the indirect paths via early smoking and adolescent drinking, a manual parental social class was only associated with a lower likelihood of young adult drinking in BCS70 (OR was 0.64 compared to 1.03 and 1.00 in NCDS and T07). Gender-stratified analyses produced very similar results.

Conclusion Early smoking may mark out mechanisms related to socioeconomic disadvantage which promote heavier drinking, but there is otherwise a tendency for the more affluent to drink more heavily, which seems consistent in adolescence, but more context dependent in young adulthood.