European Journal of Personality vol:21 issue:1 pages:63-89
The spectrum hypothesis, postulating that differences between referred and non-referred samples are confined to mean level differences, is elaborated by exploring whether the covariation between child problem behaviour and its predictors-child personality and parenting, rated by mothers-is similar in referred (N = 205) and non-referred (N = 596) children and whether personality by parenting interactions can be generalized across samples. Results showed significant mean level differences for all the variables. Both personality and parenting explained problem behaviour with some differences in strength of the effects across samples. Parenting by personality interactions mainly predicted externalizing behaviour, with benevolence and conscientiousness as the most prominent moderators. Results confirmed that moderators of problem behaviour operate similarly in the two samples, thus corroborating the spectrum hypothesis. Copyright (c) 2006 John Wiley & Sons, Ltd.