Notes to Pages 93–106- In online Table A6.2 we show, for those surveys where the data
permit, a fixed- effects analysis. - See online Table A6.2.
- See online Descriptive Statistics.
- In Australia, the results here confirm that mental health has the
largest single impact. For Britain mental health has to be entered with
a lag, and its effects are therefore understated. - Layard and D. M. Clark (2014).
- See Dolan (1997).
- In other words the average of the shaded bars has been made
equal to the average of the black bars. For the life- satisfaction regres-
sion, see Dolan and Metcalfe (2012). - To get the full QALY impact of a condition we also of course have
to add its impact on longevity, but that is not our concern in this book. - Layard and D. M. Clark (2014).
- On the negative impact of partner’s illness on the caregiver in
Australia, see van den Berg, Fiebig, and Hall (2014). - In Britain this is in measured by “is disabled”; in Germany it is
measured by “registered as disabled.” For earlier work on the BHPS, see
Oswald and Powdthavee (2008). - Dolan and Metcalfe (2012).
- Plomin et al. (2013). On the issue of genes, see Chapter 12.
- Danner, Snowdon, and Friesen (2001).
- Steptoe, Deaton, and Stone (2015).
Chapter 7. Crime
- For the full version of the paper on which this chapter is based,
see online Annex 7. - The predictive power of the equation is not huge. In Figure 1.4
the equation for crime has an R^2 of 0.10. - This is a different dependent variable for that in Figure 1.4,
where the dependent variable was the number of times arrested by the
age of 34. The convictions variable is measured at age 30 so as to facili-
tate comparison with the US data. - The controls are more limited than those in the rest of the book
to facilitate the UK/US comparison that follows. - At both ages one standard deviation extra of bad behavior raises
the probability of arrest by nearly five percentage points— a 33% in-
crease in risk. See online Annex 7.