The longitudinal, expert, all data (LEAD) procedure has been employed as a criterion for the assessment of the procedural validity of diagnostic instruments. This study evaluated the procedure's concurrent, discriminant and predictive validity. Interview and questionnaire data obtained from 100 individuals in a substance abuse treatment program were used to assess current and lifetime substance use disorders and common comorbid disorders. An experienced, doctoral-level clinician formulated LEAD diagnoses for each patient, based on an initial interview, ongoing clinical contact and the results of the research assessment and all available clinical records. LEAD-derived substance use diagnoses showed good concurrent, discriminant and predictive validity. The validity of comorbid diagnoses obtained using the LEAD procedure was generally fair to good. Comparison with diagnoses based only on the clinician's unstructured initial interview showed that the availability of additional data enhanced diagnostic validity. Diagnoses derived by a research technician using the Structured Clinical Interview for DSM-III-R showed validity comparable to that of LEAD diagnoses. To enhance its diagnostic validity, applications of the LEAD standard should include a structured interview. Other variations in the application of the LEAD standard, including a longer evaluation period, may also enhance its performance as a diagnostic criterion measure.