The name is absent

Frequency of drinking alcohol

As seen in Table 1 above, there is a substantial increase between the 1958 and the
1970 cohort in the percentage of men and women who say that they drink alcohol on
most days. This is particularly marked for women. To investigate this in more detail,
logistic regression models were estimated that took whether an individual reported
drinking on most days or not as the dichotomous dependent variable. By pooling data
from the 1958 and 1970 cohorts it was possible to look at the associations between
gender and social class and the frequency of drinking, and to see whether there were
any interactions between cohort and these two explanatory variables. Table 5a below
summarises the results of this analysis.

Table 5a : Social Class, Gender and Cohort as predictors of drinking on most
days.

It can be seen that across the whole sample including both cohorts, women were less
likely to report drinking on most days than men (with a coefficient of -0.998) and this
confirms the bi-variate analysis shown above. However, the fact that there is a
significant positive interaction between cohort and sex shows that the differences
between men and women were more marked for the 1958 cohort than for the 1970
cohort. In other words, the magnitude of the negative coefficient for women (-0.998) is
reduced by 0.331 for women in the 1970 cohort. In addition, there is a significant
relationship between social class and frequency of drinking, such that those in the

<<   13   14   15   16   17   18   19   >>

More intriguing information