Initially a set of descriptive bivariate analyses have been carried out. Where there
appear to be substantial differences between the two cohorts, the data have been
pooled to form a single dataset to enable multivariate analysis of data including
interaction terms - for example, between cohort and gender - on the outcomes of
interest.
Results
As can be seen from Table 1, the reported frequency of drinking alcohol of men in
their early thirties is similar for the 1958 cohort and the 1970 cohort. In 1991 86.5 per
cent of men reported drinking at least once a month, by 2004 this had only risen
slightly so that 87.6 per cent of men reported drinking at least once a month. The
increase was slightly more pronounced for women, rising from 71.1 per cent of
women to 74.2 per cent of women. Table 2 shows that while just over a quarter of
men reported that they drank between 0 to 3 units of alcohol each week, just over ten
per cent of men in each cohort are heavy drinkers, consuming 36 or more units of
alcohol per week. As we would expect, the results show that women tend to drink
much less than men; however a comparison between the two cohorts shows women
born in 1970 tend to drink more than their counterparts in the earlier cohort. For
example, if we focus on moderate and heavy drinkers (defined as those drinking
more than 9 units per week), among the 1958 cohort 19.4% of women were either
moderate or heavy drinkers, compared with 25.1% of women in the 1970 cohort.