children were at least two years delayed in their reading after allowing for IQ differences. These associations held even after controlling for family size and social class. There are four major hypotheses concerning the overlap between underachievement and conduct disorder: (1) underachievement leads to externalizing behavior, (2) externalizing behavior leads to underachievement, (3) hypotheses 1 and 2 occur simultaneously, and (4) underlying variables are the actual cause in both problems. All four hypotheses need further exploration to derive explanatory models with sufficient rigor, complexity, and validity to handle the diversity of causal factors (Hinshaw, 1992).
Children with chronic ill health, if they also have a functional physical limitation (any sensory, physical disorder, or disability that interferes with functioning), have been found to have three times the incidence of conduct disorder as healthy peers (Cadman, Boyle, Offord, Szatmari, Rae Grant, Crawford, and Byles, 1986). Children with chronic ill health without disability have been found to have twice the incidence of conduct disorder (Cadman et al., 1986). Chronic conditions affecting the central nervous system place the child at even higher rates of risk; children with central nervous system damage have five times the incidence of conduct disorder, according to studies by Rutter (1977) and Brown and colleagues (Brown, Chadwick, Shaffer, Rutter, and Traub, 1981). It is not known whether brain damage exerts its effect primarily through the presence of cognitive disability per se or through the resulting difficult temperament on the part of the infant.
Identification of patterns of early precursor symptoms may lead to identification of children and adolescents who are at especially high risk for developing conduct disorder. Beginning at ages four to six, aggression predicts later delinquency (Loeber and Dishion, 1983). When aggression is combined with other behavioral characteristics, however, the predictive power increases. A combination of aggressiveness and shyness appears to be predictive of conduct disorder as well as drug abuse (Farrington and West, 1990; Moskowitz and Schwartzman, 1989; McCord, 1988a,b; Kellam, Brown, Rubin, and Einsminger, 1983). Farrington and colleagues from the Cambridge Study in Delinquent Devel-