Reflections of dietary studies with children in the ten-state nutrition survey of 1968-1970.

The dietary component of the Ten-State Nutrition Survey’ (TSNS) was designed to provide a description of food consumption for persons of many ages, including infants and small children 0 to 36 months old and adolescents 10 to 16 years old. Because the overall objective of the TSNS was to describe nutritional status of “groups” as contrasted to “individuals,” the 24-hour recall method was an acceptable technique; however, its value is obviously limited by the ability of the respondent (mothers in the case of infants and young children) to remember all food items consumed. The actual range of intake within “groups” of the TSNS was probably less than that reported’ because of both underestimation and overestimation of food consumption. Average intakes of some nutrients determined from 24hour recalls may be some 10% greater than those found by other techniques, indicating that 24hour intakes are more likely to be overestimated than underestimated.2 This fact does not invalidate use of 24-hour recall data for comparing intakes of “groups” of persons included in one survey if the assumption is made that erroneous data are randomly distributed among the subsets of “groups.” Examination of the detailed reports of the TSNS1 and of raw data#{176}indicated that erroneous data were randomly distributed.