Validation of a food quantification picture book and portion sizes estimation applying perception and memory methods

Abstract

We tested the applicability of EPIC-SOFT food picture series used in the context of a Hungarian food consumption survey gathering data for exposure assessment, and investigated errors in food portion estimation resulted from the visual perception and conceptualisation-memory. Sixty-two participants in three age groups (10 to <74 years) were presented with three different portion sizes of five foods. The results were considered acceptable if the relative difference between average estimated and actual weight obtained through the perception method was ≤25%, and the relative standard deviation of the individual weight estimates was <30% after compensating the effect of potential outliers with winsorisation. Picture series for all five food items were rated acceptable. Small portion sizes were tended to be overestimated, large ones were tended to be underestimated. Portions of boiled potato and creamed spinach were all over- and underestimated, respectively. Recalling the portion sizes resulted in overestimation with larger differences (up to 60.7%).

Keywords:

• Portion size estimation
• validation
• food photographs
• perception
• conceptualisation-memory

Acknowledgements

The authors are grateful to the Budapest Business School and the staff of the institute for participating in preparation and conducting of the study; Nadia Slimani and her team for providing the know-how of the EPIC-SOFT methodology and the picture book; the PANCAKE project for providing their base documents on the validation of the picture book; and the members of PANEU consortium for their experience on validating picture book and statistical analysis.

Disclosure statement

The authors report no conflicts of interest.