European Journal of Clinical Nutrition · 5 authors, 3 centres
This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.
An online image-series tool displaying eight portion size options for 15 discretionary foods was developed and validated against equivalent real food options in a randomised-crossover laboratory study. The tool demonstrated good agreement with real food selections, with over 90% of choices matched to the same or adjacent portion size option.
This study developed and validated an online image-series tool for assessing perceived portion size norms of 15 commonly consumed discretionary foods, each with eight portion size options. A sample of 114 adults (mean age 24.8 years) completed a randomised-crossover laboratory session in which they reported perceived portion size norms twice—once using computer-based food images and once using equivalent real food options at laboratory food stations. Agreement was assessed via cross-classification and intra-class correlation (ICC). Cross-classification indicated that over 90% of selections were matched to the same or adjacent portion size option, and the overall ICC was 0.85, indicating good agreement. Three foods (banana bread, muffin, chocolate bar) showed only moderate ICCs, possibly due to wide variability in typical serving sizes or difficulty distinguishing features from two-dimensional images. Cooking confidence was positively associated with agreement levels. Limitations include a convenience sample primarily comprising young, educated females and potential carryover effects within a single session, though randomisation of presentation order mitigated the latter.
