The efficacy of cognitive—behavioural therapy (CBT) for the treatment of childhood anxiety has mainly been demonstrated in university-clinic settings. This study aimed to evaluate the effectiveness of CBT for the treatment of childhood anxiety in a community mental health service, compared with standard treatments (‘Treatment as Usual’ [TAU]) 'child psychotherapy, family therapy and eclectic treatments. Fifty-four children with anxiety disorders, aged from 7 to 14 years, were randomly assigned to either a CBT or TAU group. CBT and the standard treatments were provided by 18 experienced therapists, and the mean number of treatment sessions was 12. Baseline and follow-up measures at 3, 6 and 12 months included an interview based on criteria in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) that was conducted by a clinical psychologist who was blind to the treatment conditions, and standardised anxiety measures (parent, child and teacher reports). Significant improvements were found on all anxiety measures over time, but no significant differences were found between CBT and TAU. The challenge of conducting such research in community settings is discussed.