Different patterns of boys' externalizing behavior and their relation to risk factors: a longitudinal study of preschool children.

Childrens' externalizing behaviors such as aggression, delinquency and impulsivity are serious problems in many societies. In previous person-oriented analyses we found two types of externalizing problems in boys. One pattern contained externalizing problems only, whereas the other type showed both externalizing and internalizing problems (anxiety, depression etc.). The present study addressed these two groups in a prospective longitudinal design. It was investigated whether the previous descriptive findings remained stable over time and, in particular, whether the two types differed in important risk factors for antisocial behavior. The sample consisted of 198 boys from the Erlangen-Nuremberg Development and Prevention Study. The first assessment took place in kindergarten and the second 3.4 years later in elementary school. The behavior problems were assessed by ratings of kindergarten teachers and elementary school teachers using the Social Behavior Questionnaire (SBQ). The risk factors were low socio-economic status of the family, birth complications, physical punishment in parenting behavior, difficult temperament, low intelligence, and aggression-prone social information processing of the child. Approximately 15% of the boys revealed externalizing behavior problems. A variable-oriented analysis showed significant stability over time. In a person-oriented Configural Frequency Analysis the 'externalizing only' and the 'combined externalizing and internalizing' pattern could be replicated. For five of the six risk factors we found significant differences between the boys with behavior problems and a non-deviant group. However, the two different patterns of externalizing problems differed only in intelligence (lower for the group with combined problems). The results confirm models of cumulative biological, psychological and social risks for antisocial behavior over time. Furthermore, specified analyses of the two types and their relation to proactive and reactive aggression are recommended.