Obesity is generally associated with hyperinsulinemia. However, whether obesity precedes or follows hyperinsulinemia is not clear. The present study examined the temporal nature of the association between obesity and hyperinsulinemia in a biracial (black-white) community-based population enrolled in the Bogalusa Heart Study. Three longitudinal cohorts of children (n = 427; baseline age, 5 to 7 years), adolescents (n = 674; baseline age, 12 to 14 years), and young adults (n = 396; baseline age, 20 to 24 years) were selected retrospectively, with a follow-up period of approximately 3 years. In general, longitudinal changes in the mean body mass index (kilograms per meter squared), an indicator of adiposity, and fasting insulin level did not parallel each other. In a bivariate analysis, baseline insulin levels correlated significantly with the follow-up body mass index in adolescents and adults, but not in children. On the other hand, the baseline body mass index correlated significantly with follow-up insulin levels in all cases. Logistic regression analysis showed that the proportion of subjects who developed obesity (body mass index > 75th percentile, specific for age, race, gender, and survey year) at follow-up study increased significantly across baseline quintiles (specific for age, race, gender, and survey year) of insulin only among adolescents, irrespective of race and gender. This relationship disappeared after adjusting for the baseline body mass index. By contrast, a significant positive trend between baseline quintiles of the body mass index and incidence of hyperinsulinemia (> 75th percentile) at follow-up study was noted among all age groups independent of race, gender, and baseline insulin levels. Further, in a multiple stepwise regression model, the best predictor of the follow-up insulin level was the baseline body mass index in children and adults and the baseline insulin in adolescents. The baseline body mass index was the best predictor of the follow-up body mass index in all three age groups. These results, by showing the temporal nature of the relation between obesity and hyperinsulinemia beginning in childhood, support the role of obesity in the development of hyperinsulinemia.