that those in the manual social classes were much more likely to be heavy drinkers
than those in the non-manual social classes.
Longitudinal analyses of the links between childhood and adolescent characteristics
and drinking behaviour in early adulthood found that for women, heavy drinking at
age 23 (defined as over 35 units per week) was associated with better housing and
less financial hardship in childhood; having a smaller family of origin with a skilled
manual father; going out more at age 16; smoking at age 16; and spending more
money on entertainment and alcoholic drinks at 16 rather than on savings. Separate
analyses were carried out for men, and the results were found to be rather different,
with none of the social and family background variables distinguishing the heavy
drinkers from the other groups. However, it was found that being a heavy drinker at
age 23 for men (defined as over 50 units per week) was associated with more family
conflicts and more extroverted leisure activities in adolescence as well as being a
smoker at age 16 (Ghodsian, 1985).
Analysis of data from the 23 and 33 year surveys of the 1958 cohort (Power et al
1999) found that overall rates of heavy drinking declined substantially between ages
23 and 33 but persisted for never married men and women and increased
significantly among individuals who divorced compared with those who remained
married. Furthermore the levels of heavy drinking among these young adults in 1991
was not found to be due to selection effects and the authors concluded that marital
separation had a pronounced short-term effect on heavy drinking.
More recently, research by Jefferis, Manor and Power has used data from the 1958
cohort to examine the social gradients in binge drinking and non-drinking at different
points in the life course (Jefferis et al 2007). Their analyses demonstrate that the
least educated men reported non-drinking or binge drinking more often than more
educated men throughout adult life at ages 23, 33, and 42 years. For women the
pattern of results was somewhat different. At age 23 it was the better educated
women who were most likely to be binge drinkers, and by age 42 this trend had
reversed so that in mid-adulthood the social pattern of binge drinking in women more
closely resembled the results obtained for men.
Much of the research to date has focused on either the 1958 cohort study or the
1970 cohort study separately, with very little comparative analysis examining
differences between the drinking patterns of these two cohorts. The data now
available from the 2004 sweep of BCS70 at age 34 makes it possible for the first time
to compare the drinking behaviour of individuals in their early thirties in the two
cohorts.
Analysis
The analyses in this paper have focused primarily on the differences between the
drinking behaviour of 33 year olds in 1991 (namely those from the 1958 cohort study)
and the drinking behaviour of 34 year olds in 2004 (namely those from the 1970
cohort study). In particular the emphasis is on providing a detailed description of how
drinking patterns vary by gender and social class for these two cohorts born twelve
years apart.