Currently there are two competing methods for evaluating and diagnosing children with psychiatric disorders. There are marked differences between them. One is represented by the DSM-IV-derived structured interview and is symptom based; the other is the traditional open-ended interview, sometimes using play or projective materials. The first is quantitative and seeks accuracy; the second is more qualitative and seeks meaning. Each has its own advantages and disadvantages. Together they represent the science and the art of child and adolescent psychiatry. This article highlights the critical importance of a comprehensive and balanced formulation and offer an approach to integrate the three influences, biological, psychological, and social, into a clinically meaningful framework for treatment planning. Topics covered include: Limits of the DSM Symptom-Oriented Approach to Diagnosis; The Diagnostic Process; Biopsychosocial Linkages in the Diagnostic Process; and a Biopsychosocial Formulation for Clinical Practice.