Table 5.
Comparison of portion size estimation accuracy across the type of food presentation in the image-series validation study, the Norwegian version of myfood24
| Difference in portion size estimations | Correct estimation | P | |
|---|---|---|---|
| Median | IQR | ||
| Food presentation | |||
| As depicted (n 36) | 0⋅60 | 0⋅51 | 0⋅416 |
| Not as depicteda (n 10) | 0⋅63 | 0⋅43 | |
| Presented weight | |||
| Identical weight (n 23) | 0⋅59 | 0⋅49 | 0⋅597 |
| Altered weightb (n 23) | 0⋅61 | 0⋅59 | |
Mann–Whitney U test. Correct estimation referring to the participants’ mean proportion of correctly classified portion size image estimates per dish/food item, correctly classified defined as a perfect match between the portion size image chosen by the participant and the portion size of that same food or dish presented to the participant. Median represents the central tendency of the mean correct estimates per dish/food item, with interquartile range (IQR) representing the measure of variability.
Presented with a different plate or bowl than depicted.
Presented with ±25 % of the differential to an adjacent portion size image.