Notes to Pages 83–93- We do of course know whether some people in the sample have
had children who have left home, but we do not know this for those
who joined the sample after their children had left. - For BCS see Online Table A5.1, and for the household panels,
see Online Table A5.2. - See Myrskyla and Margolis (2014); Cetre, Clark, and Senik
(2016); online Annex 5. The numbers in Figure 5.6 include controls.
Chapter 6. Health of Mind and Body
- The anterior cingulate cortex and anterior insula; see Layard
and Clark (2014), 274n37. - Layard and D. M. Clark (2014).
- For a classic analysis of how external events can cause mental
illness, see G. W. Brown and Harris (1978), and on the general causes of
mental illness, see Layard and D. M. Clark (2014), chapter 7. - Life- satisfaction is an integer variable, so it is not possible to de-
fine the same percentage of the population as miserable in every coun-
try. In addition the number of categories of life- satisfaction is 11 in the
BCS and HILDA, 7 in the BHPS, and 4 in the BRFSS. The percentages
defined as in misery are shown in Table 6.1. - Behavioral Risk Factor Surveillance System.
- They also ask whether you are currently in treatment (which of
course understates prevalence). The questions used in all the surveys
appear in Annex 6. - The distribution of conditions is as follows (%):
0 1 2 3 4 5+
BCS 37 34 17 7 2 1
BHPS 38 30 17 8 4 3- For a simple description of the characteristics of the most mis-
erable people, see online Table A6.1. - Logit analysis produces very similar results.
- The relation between Tables 6.1 col. 3 and Table 6.2 for any dis-
continuous variable is as follows. In Table 6.1 we estimate an equation
that implies pM = ∝ipi + etc. where pM is the proportion in misery and
pi is the proportion with characteristic i. Column 3 is ∝i pi. By contrast
for each characteristic Table 6.2 shows ∝i
pi(1 – pi)
pM(1 – pM).