During all waves of the study, adolescents' depressive symptoms predicted perceived parental rejection, and perceived parental rejection predicted adolescents' aggression. Boys and girls did not significantly differ from one another in this regard, with the exception of perceived parental rejection in relation to aggression between the second to third wave. For the adolescent group as a whole and the boys and girls groups separately, the strength of the prediction of perceived parental rejection, adolescents' depressive symptoms, and aggression became weaker over the subsequent waves of data collection. The authors advise that the aforementioned effects cannot be revealed when global measures of adolescent problem behaviors are used (e.g., Conger and Simons, 1997; Spoth et al., 2006) or when the study is a cross-sectional or a short-term longitudinal design (e.g., Heidgerken et al., 2004; Newman et al., 2007). Over a 4-year period, data were collected yearly from 940 early adolescents (50.6-percent boys and 49.4-percent girls), who completed self-report questionnaires that measured depressive symptoms, aggressive behaviors, and perceived parental rejection at 1-year intervals. The adolescents were drawn from 12 different Dutch junior high and high schools in the Utrecht Province of The Netherlands. Depressive symptoms were measured with the Children's Depression Inventory. Aggression was measured with the Direct-Indirect Aggression Scale (Bjorkqvist et al., 1992); and perceived parental rejection was measured with the Level of Expressed Emotion questionnaire. Structural equation modeling based on maximum likelihood estimation was performed in AMOS (Arbuckle and Wothke, 2006). 2 figures, 4 tables and 50 references