340
THE ECONOMIC AND SOCIAL REVIEW
If we use the narrower definition of free-time, which is confined to time
spent on leisure plus religious/civic activity, the 60 per cent median thresholds
are lower (in terms of hours). On weekdays the time poor are those with
less than 2 hours 42 minutes of leisure time and on weekend days it is those
with less than 4 hours of leisure time. On this measure 20 per cent of respond-
ents are defined as time poor on weekdays and 22 per cent on weekend days.
Looking at the characteristics of those who are time poor we find that the
patterns are broadly similar to those emerging when we examined average
time-use figures. Women are more likely to be time poor at weekends but the
difference during weekdays is not statistically significant. Employment at the
individual and household level continues to be important. On the
uncommitted time measure 8 per cent of the self-employed are time poor on
weekdays in contrast with the retired, none of whom are time poor. Similarly,
on the narrower leisure based measure the self-employed have the highest
risk of time poverty on weekdays and joint highest risk (along with those in
home duties) on weekend days. Dual-earner couples and male-breadwinner
couples show similar levels of time poverty on weekdays using both measures,
and at weekends using the leisure based measure, but dual-earners are less
likely to be counted as time poor at weekends using uncommitted time. Time
poverty is particularly pronounced among parents of young children. On
weekdays, 15 per cent of those with pre-school children are defined as ‘time
poor’in terms of uncommitted time and almost half (47 per cent) are time poor
in terms of leisure time. This group also have the highest risk of time poverty
at weekends. Children of any age under 18 years increase the risk of time
poverty.
The relationship between time poverty and household income is not
linear. Those in the highest income quartile experience the highest rate of
time poverty using the leisure measure, which is consistent with the
hypothesis that ‘busyness’ is higher amongst the wealthier, but this does not
hold true for uncommitted time at weekends. A similar pattern emerges for
education. Those with the highest education are at greatest risk of time
poverty on three of the measures (not for uncommitted time at weekends) but
the relationship between time poverty and the other education levels is non-
linear.
3.4 Models of Committed Time
The descriptive tables outlined above do not allow us to establish the
independent effect of different factors on peoples’ total workload. Therefore,
we construct OLS regressions of committed time for both weekends (Table 9)
and weekdays (Table 10). When factors such as employment status and family
status are controlled a significant gender difference in committed time