Influence of energy density and unit size on portion size estimation

Abstract

Consumption of large portion sizes has been associated with positive energy balance and in the long-term, obesity. How much is appropriate to consume may be overestimated for energy dense foods and for snacks/meals containing multiple items. This study tested the hypothesis that energy density (ED), unit size and meal type influences portion size estimation. Thirty-two lean adults estimated the number of portions presented of 33 foods. Portions displayed corresponded to 0.5–5.3 FSA reference portions (1.7–26.8 kJ/g), and included 12 multi-item and 21 single unit foods; 13 were labelled “meal”, 4 “drink”, 16 “snack”. The main study outcome was percent error of estimation relative to the reference portion. The Wilcoxon test was used to compare portion estimates against reference portions and %error of estimation between snacks vs. non-snacks, and unit vs. multi-item foods using paired data. We tested the association between %error and ED with Spearman’s rank correlation. The median portion estimate across foods was 1.2 (0.8–2.9). Participants overestimated the size of portions corresponding to ⩽1 reference amount (9 foods) and underestimated it for portions >1 reference amount (19 foods). Percent error was not influenced by unit size or meal classification, and was not associated with ED in the whole food sample, nor in snacks, meals/beverages or single unit foods, but correlated positively with ED in multi-item foods ( r = 0.69; p