Aggression and mental ill-health in adolescent friendship groups : A 4-year longitudinal study

Friends can have a powerful influence on each other’s behaviours (Brown & Klute, 2003; Brown & Larson, 2009). Investigation of the influence of friends tends to focus either on externalising behaviours, such as aggression, or on internalising behaviours, such as mental ill-health. To date, these two lines of research have remained largely separate from each other with little or no research devoted to the investigation of both aggression and mental ill-health in the context of friendship groups. This is surprising given the well-established association between individual aggression and individual mental ill-health. Thus, the main aim of this thesis was to investigate the links between aggression and mental ill-health over time in the context of adolescents’ friendship groups. A further aim was to make a methodological contribution to this area of research by a comparison of the results obtained when different friendship conceptualisations (reciprocal friends, non-overlapping friendship groups, and overlapping friendship groups) are used to investigate the substantive questions related to the associations between aggression and mental ill-health. Data for this thesis came from 2,865 high school students (50.44% female) from 17 Catholic schools in Australia. Longitudinal data were collected yearly from Grade 8 to Grade 11 to measure aggression, mental ill-health, and peer nominated friendships. Multilevel analyses, presented in Study 1, indicated that high levels of individual aggression predicted high levels of friendship group aggression over time. A similar result was found for mental ill-health, whereby high levels of individual mental ill-health predicted subsequent high levels of friendship group mental ill-health. Furthermore, when individuals engaged in aggressive behaviours, they tended to suffer from worse mental ill-health over time. Interestingly, a homeostatic process emerged between the friendship group variables, whereby high levels of group aggression predicted subsequent high levels of group mental ill-health, and high levels of group mental ill-health predicted subsequent lower levels of group aggression. Notably these results differed somewhat depending on whether the friendship groups were conceptualised as being reciprocal, non-overlapping, or overlapping, suggesting that peer influence processes operate differently across different conceptualisations of friendship. Problematic levels of collinearity observed between group aggression and group mental ill-health, when the overlapping friendship conceptualisation was used, prompted further analyses, presented in Study 2. Specifically, I investigated how attributing different group weights impacted the relations between aggression and mental ill-health at both the individual and the group level. The results of this study demonstrated the importance of selecting a theoretically driven weighting metric when analysing overlapping friendship groups, with an individual’s most influential group being the most appropriate for achieving the research aims of this thesis. Finally, in Study 3, I sought to provide further insight into the substantive findings of this thesis by investigating the role of both selection and socialisation processes in the associations between individual aggression and group aggression, and between individual mental ill-health and group mental ill-health. The results supported the simultaneous effects of selection and socialisation processes for aggression. Moreover, socialisation processes, but not selection processes, were observed for mental ill-health. In summary, investigation of both aggression and mental ill-health in the context of three different friendship group conceptualisations has enabled unique insights into, and further understanding of, the developmental psychology field.