Chapter 7the probability of conviction, other things equal. The esti-
mated coefficients are in the second column of Table 7.1
and (allowing for the greater levels of crime in the US) are
broadly similar to those found in Britain. When childhood
behavior is worse by one standard deviation, this raises the
probability of conviction by 6 percentage points. That is an
increase of 28% over the average conviction rate— almost
exactly the same as in Britain.
Does Education Reduce Crime?
If early childhood problems tend to produce more crime, can
extra years of education offset this? Indeed, is one important
by- product of education a reduction in crime? Our BCS data
provide a first approach to this issue. In Table 7.2 we regress the
number of convictions by age 30 on qualifications obtained, as
well as on childhood outcomes at age 10 and family variables.
Table 7.2. How the number of convictions by age 30 is affected by
qualifications, childhood outcomes at 10, and family background
(British Cohort Study)UnitsQualifications SD (index) −0.06 (.01)
Intellectual performance (age 10) SD (index) −0.01 (.01)
Good behavior (age 10) SD (index) −0.13 (.02)
Emotional health (age 10) SD (index) 0.07 (.01)
Family characteristics SD (index) −0.17 (.03)