BACKGROUND: This study examined the degree of overlap and sources of discrepancies between two alternative models of social cognition in early adolescence, a general model based on 'theory of mind' and a relationship-specific model rooted in attachment theory. METHODS: We administered newly developed, age-appropriate measures of advanced theory of mind or 'mentalising' and child-parent attachment, indexed by the coherence of children's narratives about attachment relationships, to a sample of 70 early adolescents (mean age 12.6 years). RESULTS: The central findings were that, as expected, there was significant overlap between the mentalising and attachment indices of social cognition (r(70) = .35, p < .01). The overlap was not significantly mediated by verbal ability or general measures of parenting. There were also substantial discrepancies in children's performance on these measures that were conceptually interesting and statistically reliable. Specifically, adolescent reports of parenting discriminated between measures of mentalising and attachment coherence. In addition, those children who exhibited a less coherent model of attachment than was predicted from performance on a mentalising task were more likely to be rated as exhibiting a 'Dismissing/Avoidant' style in the attachment interview. CONCLUSIONS: Findings support a connection between mentalising and attachment in early adolescence, and further suggest that social cognitive processes may be context and relationship-specific.
BACKGROUND: This study examined the degree of overlap and sources of discrepancies between two alternative models of social cognition in early adolescence, a general model based on 'theory of mind' and a relationship-specific model rooted in attachment theory. METHODS: We administered newly developed, age-appropriate measures of advanced theory of mind or 'mentalising' and child-parent attachment, indexed by the coherence of children's narratives about attachment relationships, to a sample of 70 early adolescents (mean age 12.6 years). RESULTS: The central findings were that, as expected, there was significant overlap between the mentalising and attachment indices of social cognition (r(70) = .35, p < .01). The overlap was not significantly mediated by verbal ability or general measures of parenting. There were also substantial discrepancies in children's performance on these measures that were conceptually interesting and statistically reliable. Specifically, adolescent reports of parenting discriminated between measures of mentalising and attachment coherence. In addition, those children who exhibited a less coherent model of attachment than was predicted from performance on a mentalising task were more likely to be rated as exhibiting a 'Dismissing/Avoidant' style in the attachment interview. CONCLUSIONS: Findings support a connection between mentalising and attachment in early adolescence, and further suggest that social cognitive processes may be context and relationship-specific.