Estimation of the population mean for correlated observations. Applications to Nutritional Epidemiology

The procedures for estimating intakes of both energy and different nutrients in individuals and populations are imperfect because they do not consider the correlation between intakes in the coming days or the fact that individuals have different patterns of variability of intakes. Random-effects statistical models, which are those used in this modeling, usually do not consider correlations of this type either, only those that derive from repetitions of observations being made to the same individual. It is necessary to modify Beaton's model to this more realistic situation, to determine analytically how correlations and heterocedasticity affect the process of estimating intakes of individuals and populations, as well as to determine the procedures for estimating the population distribution of the number of days needed to evaluate intakes.

Proyecto cancelado por desvinculación del docente