Mechanisms and Extrapolation in the Abortion-Crime Controversy - Mechanism and Causality in Biology and Economics - page 180

psychological and social development, while the second has to do with factors that
would make the pregnancy unwanted in the first place. To illustrate the first of these
two sub-mechanisms, a mother might resent a child resulting from an unwanted
pregnancy and thus have a more negative and less affectionate attitude towards it,
which could exert a damaging psychological effect on the child. To illustrate the
second, the mother might desire to terminate the pregnancy because she is an
unmarried teenager, and children born to unwed teens, regardless of whether they
are wanted or unwanted, may be at greater risk for a number of adverse life
outcomes, including criminality.
The most direct empirical test of the selection mechanism would identify
children whose mothers desired to terminate their pregnancies but were prevented
from doing so by legal restrictions on abortion. Data sets of this kind exist for
several European countries wherein, during the 1930s-1960s, women desiring to
terminate a pregnancy were required to file an application to obtain legal permis-
sion to do so. Thus, cases of women whose applications for abortion were denied
and who subsequently gave birth constitute precisely the type of sample in question.
Samples of this kind exist for Sweden and the former Czechoslovakia and have
been the basis for a number of studies documenting the psychological and social
effects on children of being born unwanted (see David et al. 1988 ). The most
thorough of these studies concerns the Prague cohort of 220 children born to
women whose request for abortion was twice denied, on the initial application
and then on appeal. This cohort is known as UP (for unwanted pregnancy) and was
matched with 220 AP (accepted pregnancy) control children. Since the Prague
study was designed to test of the effects on the child of being unwanted (i.e., the first
sub-mechanism mentioned above), the UP and AP subjects were matched on
socioeconomic terms as well as on a number of other factors such as family size.
In addition to collecting medical, school, and legal data, the study conducted
double-blind interviews of parents, teachers, and children, which for the latter
group included psychological and intelligence tests. The study focused on several
age points: birth, age 9, ages 14-16, and ages 21-23. The study found no physio-
logical or health differences between the UP and AP children at birth but a
consistent pattern of less favorable outcomes in the subsequent follow-ups. For
example, at age 9, UP children were significantly more likely to be rejected by peers
(Matejcek et al. 1988 , pp. 69-70) and to have difficulty in adapting adequately to
frustration (ibid. 70-71). By the age 14-16 follow-up, the gap in school achieve-
ment between UP and AP children had become statistically significant (ibid. 88).
By the age 21-23 follow-up, more than twice as many UP subjects had been
sentenced in court, and the average prison term of those sentences for the UP
subjects was more than double the average sentence of the AP controls (Dytrych
et al. 1988 , p. 94). Since UP and AP families were matched for socioeconomic
status, most of these differences appear to be due to the less favorable internal
family dynamics of the UP subjects (Matejcek et al. 1988 , pp. 74-75).
The European studies are cited as support for the selection mechanism in both
the original Donohue and Levitt article ( 2001 , p. 388) as well as in a summary of
that argument given by Levitt ( 2004 , p. 182). The importance of these studies in