Chapter 9well- being of older men and women. We have therefore
tested how changes in economic, social, and personal fac-
tors at older age relate to trajectories of life- satisfaction.
We explored these possibilities by studying changes be-
tween 2004 and 2012, an eight- year period, in a sample of
3,230 (55% women). The outcome variable in these analyses
was the difference between life- satisfaction measured at these
two time points, so positive scores indicate an improvement
in life- satisfaction. On the scale from 0 to 10, the average life-
satisfaction score was 7.28 in 2004 and 6.97 in 2012, so there
was a mild decrease over time on average. The changes in
characteristics considered as potential factors influencing
trajectories of life- satisfaction are described in Table 9.4. The
mean change being often quite close to zero, we also present
the percentage of people improving (going up) and getting
worse (going down), showing that there is substantial move-
ment. For example, a quarter of respondents show a shrink-
age in their social networks, while for 20% their networks
increase in size. Self- rated hearing or eyesight improve for
21% and 26% of participants respectively, which may be due
to starting using a hearing aid or spectacles.
In these analyses, our base model includes not only fixed
factors like age in 2004 and sex, but changes in income
and employment status. Interestingly, we do not observe
any differences in changes in life- satisfaction among peo-
ple who retired or moved out of paid work compared with
those whose situation remained constant; this is probably
because retirement can have both positive and negative ef-
fects, depending on the individual’s circumstances and the
measures that are used.^9
Figure 9.3 and Table 9.5 present estimates of the effects
of four sets of factors corresponding broadly to those we