The goals of the study were to examine test–retest reliability, informant agreement and convergent and discriminant validity of nine DSM–IV–TR psychiatric disorders classified by parent and youth versions of the Mini International Neuropsychiatric Interview for Children and Adolescents (MINI-KID). Using samples drawn from the general population and child mental health outpatient clinics, 283 youth aged 9 to 18 years and their parents separately completed the MINI-KID with trained lay interviewers on two occasions 7 to 14 days apart. Test–retest reliability estimates based on kappa (κ) went from 0.33 to 0.79 across disorders, samples and informants. Parent–youth agreement on disorders was low (average κ = 0.20). Confirmatory factor analysis provided evidence supporting convergent and discriminant validity. The MINI-KID disorder classifications yielded estimates of test–retest reliability and validity comparable to other standardized diagnostic interviews in both general population and clinic samples. These findings, in addition to the brevity and low administration cost, make the MINI-KID a good candidate for use in epidemiological research and clinical practice.