The name is absent

whereas men in the 1958 cohort study who were interviewed at age 33 reported
drinking an average of 25.1 units of alcohol at age 23, those who were not
subsequently included in the age 33 sweep reported drinking an average of 27.3 units
of alcohol.

A T-test was used to analyse data from each sweep to ascertain whether there are
significant differences between the group of individuals successfully interviewed at the
next sweep and the group missing at the next sweep. Figures in Table 8a and Table
8b have been highlighted in bold where the differences reach the 0.05 level of
significance. In summary, it can be seen that for men in both the 1958 and 1970
cohorts there is a tendency for heavy drinkers to be less likely to be successfully
interviewed at the next sweep. The only difference which does not reach significance
for male cohort members is for the data collected from the postal survey at age 26. In
all other cases, those who are successfully interviewed at the next sweep report
drinking approximately two units less on average than those missing from the next
sweep. The results for women are less clear cut. There is a tendency for those
successfully interviewed at the next sweep to report drinking fewer units of alcohol per
week than those missing at the next sweep. However, this only reaches significance
for the data collected from the 1958 cohort at age 42.

Table 8a: Reported mean number of alcohol units consumed per week (1958
cohort) disaggregated by whether the cohort member is interviewed in the next
sweep

Table 8b: Reported mean number of alcohol units consumed per week
(1970cohort) disaggregated by whether the cohort member is interviewed in the
next sweep

(1) Note that this will include cases missing at the next sweep due to death,
immigration, non-contact and refusal

21

<<   <   19   20   21   22   23   24   25   >   >>