Insecure attachment styles in infants are linked to mental health difficulties in later life (Fearon, Lapsley, Bakermans-Kranenburg, van Ijzendoom & Roisman, 2010), and are consequently important for identifying individuals at risk for psychopathology. The reliable and valid measurement of attachment security in infants is well documented (Ainsworth, 1979) as are measures for adults (George, Kaplan, & Main, 1985). However, reliable and valid measurement of attachment in middle childhood remains underdeveloped. The Child Attachment Questionnaire (CASQ) and Child Attachment Style Interview (CAS!) (Bifulco, Moran, Ball & Bernazzani, 2002a; Bifulco, Moran, Ball & Lillie, 2002b) were investigated for reliability and validity. Rates of attachment styles were expected to reflect those found in other studies of attachment in middle childhood. Insecure attachment identified with the CASQ was expected to correlate with higher clinical symptom scores on the Strengths and Difficulties Questionnaire (SDQ) and lower self-esteem scores on the Rosenberg self-esteem scale. 132 participants aged 9-11 years from a community sample completed the CASQ, SDQ and Rosenberg scale, of which 23 participants also completed the CASI. Rates of attachment security were similar to comparison studies in middle childhood. Children categorised with insecure attachments (both avoidant and anxious) were found to have significantly higher rates of clinical symptoms than securely categorised participants, and participants categorised as anxiously attached had significantly lower rates of self-esteem than securely categorised participants. The CASQ scales of attachment correlated significantly positively with the CASl scales of attachment. However, the overall categorisation of attachment style between the two measures was non-significant. The CASQ demonstrated significant test-retest reliability when read-ministered after five months. Evidence was found for validity and reliability of the CASQ, and validity of the CAST. High numbers were also identified in the "mixed" and "can't classify" groups. Developments of the measures and clinical implications were discussed.