Andrew M. Jones 571the introduction of HAART in 1996. There is no association pre-1996, but there is
post-1996.
12.2.3 Natural controls
12.2.3.1 Families
Auld and Sidhu (2005) present evidence that around a quarter of the association
between schooling and health is attributable to variation in cognitive ability and
that the causal effect of schooling on health is concentrated among those with
low levels of education. Estimates that allow for both schooling and health to
be influenced by a common “third factor” diminish the effect of schooling on
health except for those individuals with no greater than high school education.
The models are estimated using the 1979 and 2000 US National Longitudinal Sur-
vey of Youth (NLSY). The validity of the estimated causal effects relies on the
use of parental education as a source of independent variation, using variation
in the individual’s own education that is associated with their parents’ educational
achievements.
Siblings who are brought up together share common background characteris-
tics which may be unobserved and also influence the treatments and outcomes
of interest to researchers. Using within-sibling variation can control for these fac-
tors. Holmlund (2005) shows how variation within biological sisters can be used to
assess the long-term consequences of teenage pregnancy for educational outcomes.
The siblings approach and standard cross-section methods produce similar results
providing heterogeneity within the family is controlled for. The potential for selec-
tion bias is that teenage mothers may have family backgrounds that would lead
to poorer outcomes irrespective of an early pregnancy. Variation within biologi-
cal sisters can be used to control for these “family effects.” However, within-sibling
variation will not deal with heterogeneity within the family and the study controls
for observable pre-motherhood school performance, measured by the grade point
average (GPA) from primary school, to try and control for this. Data are taken from
a 20% sample of each cohort born in Sweden between 1974 and 1977, with the
population register used to identify siblings.
Sibling fixed effects play a role in Currie and Stabile’s (2006) study of the impact
of Attention Deficit Hyperactivity Disorder (ADHD) on educational outcomes.
Within-sibling variation is used to control for omitted variables at the level of
the family. Data from the Canadian National Longitudinal Survey of Children and
Youth (NLSCY) and the US NLSY are used. ADHD symptoms are based on parental
reports and are recorded in 1994 in the Canadian data and between 1990 and
1994 in the US data. Educational outcomes include the repetition of grades, enroll-
ment in special education, reading and math tests and delinquency. These are
measured in 1998 for Canada and 1998–2000 for the US. The study finds large
effects, relative to chronic physical conditions, and for low levels of ADHD symp-
toms in cases that would not usually receive treatment. The results for Canada and
the US are similar to each other.