Comparison of visual estimates of children's portion sizes under both shared-plate and individual-plate conditions

Objective: This paper compares the accuracy of visual estimations of children's food intake in settings where several children eat together off 1 plate vs individual-plate eating scenarios. Design: Eight trained observers were tested in their ability to estimate food portions consumed by children enacting common eating scenarios. Foods were categorized by food group and according to their presentation by individualplate and shared-plate. Observed food weight estimates were compared to actual weights. Subjects/setting: The 8 observers visually estimated 69 food portions of children eating alone and 26 portions where children were eating from a shared plate. This study was carried out in Sarlahi District, a rural, central lowland region of Nepal. Statistical analyses: Pearson's correlation coefficients were calculated to examine associations between estimated and actual weights. A fixed effects model was constructed to compare observers. Results: Analyses revealed that observer estimates of food weights under field conditions were well correlated with actual weights for individual-plate (r=0.89) and for sharedplate (r=0.84) scenarios. Observers estimated food weights when children ate together on a shared plate less accurately than they did in settings where children ate alone. With the exception of 1 observer, observers did not differ significantly in their ability to estimate food weights. Accuracy of estimations was influenced by food weight with greater error associated with food quantities of less than 70 g. Conclusions: Visual estimation is a relatively accurate, valid method of assessing child food intake under rural field conditions, and the only method to obtain accurate information on dietary intake in regions where shared-plate eating is frequent.