Measurement is a base of science. No progress exists in science, without measurement. However, measurement methods on diets, i.e., dietary assessment methods, have not been fully studied in Japan. Dietary assessment methods are science much more complicated and difficult ones than we have long believed and expected. In dietary assessments, there are many factors which induce measurement errors, both randomly and systematically. Two of most important factors are day-to-day variation of diets and underreporting of diets. "Validity" of a dietary assessment method is one of the important information which shows us how we use it and how much we can believe the data obtained from it. We, all researchers who are interested in diets, should be very careful for the validity of dietary assessment methods and the high-quality validation studies should be more encouraged.