Interest in understanding and addressing serious behavior problems in young children has increased substantially in light of growing evidence that recidivist offending in adolescence and adulthood, as well as persistent patterns of aggression and peer rejection during the early and middle school years, have their roots in disruptive behavior that can be detected as early as age 3 (Campbell et al., 1986a; Olson and Hoza, 1993; Rutter et al., 1998). Some scientists who study this issue point to evidence that antisocial behavior with a very early age of onset, compared with antisocial behavior that arises in the adolescent years, is more likely to persist into adolescence and adulthood (Caspi and Moffitt, 1995; Maughan and Rutter, 1998; Moffitt, 1997). For boys, early-onset conduct problems are moderately predictive of such adolescent outcomes as drug abuse, depression, juvenile delinquency, and school dropout (Campbell, 1991; Campbell et al., 1986b; Egeland et al., 1990; Rose et al., 1989; Wadsworth, 1976; White et al., 1990). The evidence on early-onset delinquency is, however, a matter of active debate within the field (see Loeber and Hay, 1997).
Deciding when to worry and who to worry about is not a simple matter. It is much easier to look back and say “he was always getting in trouble” or “he's always been a loner” than to predict the future trajectories of children who always seem to be getting in trouble. It is uncertain whether serious and enduring conduct problems can be predicted during the preschool years, and with what reliability. As soon as children begin to interact with one another, they begin to dislike children who hurt them. But we don't know whether or when the factors that get very young children in trouble with their peers begin to constrain the pathways they walk on the way to adulthood. What is fairly clear is that beginning in the preschool years, the social reasoning of rejected children, the skill or lack of it they display in social interaction, their ability to control their behaviors and emotional outbursts, and the nature of their interactions and relationships with adults (in particular parents) do differ from their peers in ways that are similar to differences noted for older children (Rubin et al., 1998). Thus, even if prediction from the preschool years may be tenuous, rejected-aggressive children seem to have a toehold on the pathway to later problems.
Furthermore, although rejected-aggressive children are just as likely as popular children to tell researchers that they are competent, they view others as mean, unkind, and hostile (Crick and Dodge, 1994; Dodge and Frame, 1982). This may explain why, in a recent study (Megan Gunnar, University of Minnesota, unpublished data), all of the rejected-aggressive 3-to 5-year-olds who were examined had stress hormone levels in the top third of those shown by children in the classroom. On many days the