Methodological considerations in pediatric behavioral research: measurement.

Measurement is the critical building block underlying all science. Adequate measures should demonstrate reliability, validity, and sensitivity. Special problems associated with the selection and development of measures used in behavioral pediatric research are reviewed. These include measurement specificity, construct complexity, the illusive "gold standard," construct independence, theoretical clarity, timing of measurements, and developmental considerations. The advantages and limitations of a variety of behavioral assessment strategies are described: questionnaires, interviews, behavioral observations, permanent products, biological measures, and provider ratings. Programmatic measurement research not only improves the science of behavioral pediatrics but also has widespread clinical application.