are not the reason that science is usually done; intellectual curiosity is reason
enough. The biosocial approach goes beyond the macro approaches of the past
that targeted "strain," "peer groups," "racism," "social class," "urbanism,"
"subcultures," and so forth to try to determine the mechanisms by which these
alleged causes produce their effects. However, just as I believe that any examination of criminality that does not include neurohormonal factors is inherently incomplete, any examination of neurohormonal factors that ignores
environmental factors is likewise incomplete. In this "decade of the brain,"
and with genetic knowledge increasing almost exponentially, criminologists
have an unprecedented opportunity to join other scientists in rigorous biosocial analyses of the causes of criminality. If criminologists pass up the opportunity, the torch will be passed to other disciplines; the study of crime is
too important to remain mired in premodern science.
Durkheim ( 1982:106) was careful to distinguish between social facts and biological and psychological facts, and between crime and criminality: "From the fact
that crime is a phenomenon of normal sociology, it does not follow that the criminal
is an individual normally constituted from the biological and psychological points of view."
Hirschi ( 1990:88) eschew the concept of criminality because
"It suggests that people differ to the extent to which they are compelled to crime."
The concept of criminality suggests no such thing to me, nor probably to the majority
of criminologists. Nevertheless, Gottfredson and
Hirschi ( 1990:88) wish to substitute
the concept of self-control because this concept "suggests that people differ in which
they are restrained from criminal act." Surely "criminality," stripped of the compulsive properties Gonfredson and Hirschi have imbued it with, more fully and accurately
describes the phenomenon of frequent law breaking than does "self-control," although
low self-control is a very large component of criminality.
Differential IQ also helps us to understand the underlying mechanisms of one
of criminology's most revered theories. Rowe and
Osgood ( 1984) showed that many
of the principles of differentially association theory were underpinned by differential
IQ. They found a correlation of 0.51 between self-reported delinquency and association with delinquent peers among 265 pairs of twins. This respectable correlation
should not surprise--the concept of active CE correlation informs us that people tend
to do what they enjoy in the company of the like-minded. Rowe and Osgood's data
allowed them to decompose the correlation into genetic (61%), common environment
(23%), and specific environment (16%) components. The genetic component centered
around IQ: "When twin partners have different IQs, they can be expected to have dissimilar rates of delinquency and dissimilar degrees of association with delinquent
peers" ( 1984:537). This selective aggregation process, however, may not apply in environments of greater population density where "good" and "bad" kids occupy the
same streets and buildings. In such environments, companion choice is a luxury unavailable to most children ( Stark, 1987:896).
Another problem of the same general sort exists when pooling IQ scores of
childhood-onset delinquents with adolescence-onset offenders ( Moffitt, 1993:629).
Comparing adolescent-onset and childhood-onset delinquents with nondelinquents on