Young people - particularly women - are more likely to drink heavily if they went to university or became a smoker in their teenage years, a study shows.
The discovery shows the gateway to alcohol abuse is more varied than previously believed - explaining the inconsistencies of earlier research.
A study across England, Wales and Scotland found those from a more deprived background were more likely to smoke and less likely to enter higher education.
Adolescent smokers were also more likely to drink weekly as an adolescent and more heavily in early adulthood.
But those who went to college were also more likely to drink heavily in early adulthood - with the phenomenon most prevalent among females.
More independence - free from parental control - combined with being among others of the same age group could be to blame, said the researchers.
Dr Michael Green, of Glasgow University, said: "What this study shows are the different pathways - smoking and higher education - into heavy drinking, depending on young people's socioeconomic backgrounds.
"These opposing pathways might help explain why previous research on inequalities in young people's drinking has had inconsistent results.
"It appears heavy drinking in early adulthood is more likely for both adolescent smokers and those who go to university or college.
"That would suggest the pathways to heavy drinking are more varied and opposing than had been previously thought."
The study of about 30,000 people examined in greater detail how adolescent smoking and higher education relate to the link between socio-economic background and alcohol consumption.
It measured higher education as being in full-time education after the age of 18 years old while heavy drinking was measured as more than 14 units a week for women or 21 for men - about a bottle-and-a-half of wine or seven pints of beer respectively.
The researchers believe the study may have implications for how drinking concerns are targeted and tackled in young people.
Dr Green said: "Currently interventions focused only on heavy drinking in universities/colleges are targeting a more advantaged population and may neglect more disadvantaged drinkers.
"There may be common causes affecting disadvantaged young people that lead to both smoking and heavy drinking. If we can identify and understand these it may be easier to intervene to prevent both."
Participants were from the 1958 National Child Development Study, the British birth cohort study and the West of Scotland Twenty-07 1970s cohort.
Dr Green said: "In Britain, young adults are more likely to drink heavily both if they smoke and participate in tertiary education (college and university) despite socio-economic back- ground being associated in opposite directions with these risk factors.
"In all cohorts, socio-economic disadvantage was associated with higher chances of smoking in adolescence, and adolescent smoking was associated with heavier drinking
in adolescence and early adulthood.
"However, disadvantaged adolescents were less likely to participate in tertiary education, and tertiary education was also associated with heavier drinking in early adulthood, especially for females."
He said it's presumably not the actual education but experiences associated with it that account for this.
Dr Green said: "Drinking may be a coping response to transitional challenges, may be valued for social goals or overestimation of how much peers drink may inflate perceived behavioural norms.
"However, these factors may also apply to those transitioning into work and other adult
"Perhaps increasing independence and freedom, combined with low parental monitoring, few adult responsibilities and close involvement with peers in similar situations, contribute to students' higher drinking levels.
"This association was stronger for females than males. Heavy drinking rates among UK men and women aged 16-24 converged during the 1990s, so the stronger effect of
education for females could be historical.
"If not, it could be increasingly important, as female participation in tertiary education has increased in more recent cohorts."
The study published in Addiction was funded by the Medical Research Council (MRC) and the Scottish Government Chief Scientist Office.