Objective: To determine rates of chronic physical conditions for children and youths. Design: Secondary analysis of the Health Examination Surveys, cycles 2 and 3, National Center for Health Statistics. Setting: National household survey of noninstitutionalized civilians in the United States. Participants: Multistage probability samples of children aged 6 to 11 years (cycle 2, 1963 to 1965) and youths aged 12 to 17 years (cycle 3, 1966 to 1970), and a longitudinal subsample assessed at both times. Excludes persons with an IQ lower than 80. Results: Selected items from developmental and medical histories and screening physical examinations were used to classify those with chronic physical conditions. Rates of persons with chronic conditions are 25.2% for the children and 35.3% for the youths. Parental histories produce rates equal to or greater than 11%, and physical examinations produce rates of 15.7% for children and 22% for youths. About 4% in each cycle are identified by both screening physical examination and parental history. Little overlap occurs in identification by the two sources, accounting for the higher rates for the combined indicators. Combining data sources identifies 13% of the longitudinal subsample as having a chronic physical condition at both times. Conclusions: Rates of chronic physical conditions seem to be altered by combining different indicators to categorize children. For either source used alone (eg, parental report or physician screening physical examination) the estimate is similar in magnitude, but the composition of the affected group differs. These findings have major implications for research, service, and policy. (Arch Pediatr Adolesc Med. 1995;149:1105-1110), The percentage of children and youths judged to have a chronic health condition appears to vary according to whether doctors or parents are the source of the data. Researchers analyzed data from a survey of children aged 6 to 11 years and a subsequent survey of youths aged 12 to 17. Almost one-third of the children appeared in both surveys, which allowed tracking over time. In the earlier survey, parents identified 12% of children as having a chronic condition whereas a screening physical identified 16%. Among youths, parents reported 11% as having a chronic condition versus 22% based on physical exam. For children appearing in both surveys, differing percentages of children were reported as well or ill in both time periods or as moving from one category to the other depending on whether the information source was parental report or a physical. This discrepancy has important implications for analyzing and comparing research and for determining health care policies.