This study indicated that measurements of general aggression, gathered in early adolescence, are able to predict a serious adult personality disorder, ASPD, in young adulthood in both sexes within a population-based sample. Most odds ratios in single-informant aggression models were greater than 1.5, suggesting that for those informants’ ratings of aggression (at the participant’s age at collection), each standard deviation unit increase in observed aggressive behavior corresponded to at least a 50% increased odds of having an ASPD diagnosis in young adulthood. AUC values also indicate modest levels of predictive utility. Furthermore, the results were similar regarding the direct aggression subtype, whereas there was less evidence for an independent effect of indirect aggression on ASPD. Aggression ratings from all informants significantly predicted later ASPD. However, when multiple informants were included in the model, the prediction accuracy improved.
