The development of new methods and approaches to dietary assessment has not kept pace with advances in related disciplines. One of the main gaps is the lack of innovation in incorporating cognitive psychology and computer game technology in developing useful, less biased tools for epidemiologic research on diet and disease. Another problem is the general lack of validation studies imbedded within research projects to provide data for assessment and correction of measurement error. Conceptual clarity, not only about the dietary components of interest but also about the relevant time frame for the disease under study, is often deficient in the designs of nutritional epidemiologic studies. Inadequate training of nutritional epidemiologists regarding the information requirements of a given study is also a problem, resulting either in ''overkill,'' ie, the collection of unnecessarily detailed information at great cost, or in powerless studies, where the information is inadequate for testing the hypothesis of interest.