For some foods, subjects may be able to report portion sizes relatively easily, e.g. the number and size of slice for bread. An average value could then be used for each slice size. For some foods, such as eggs or apples, questions could be included about the number eaten per week. In this case an average value could be used for one apple or one egg.
For other foods, such as meats, rice and pasta which do not come in "units", there are a number of options. An average portion weight could be used or one could assess portion sizes by means of descriptions, photographs or food models. However, portion sizes for many foods vary considerably within an individual over time and this variation has been found to be greater than that between persons of similar age and sex.
Studies investigating the effect of portion size estimates on the ability to rank individuals by their nutrient intake have shown very high correlations (r=0.9) between intake estimates using average portion sizes and those using assessed or weighed portion sizes. It has also been reported that most of the variation in intake of any food is explained by frequency of consumption. The use of average portion sizes is therefore likely to have a minimal effect on the nutrient intake estimates from a food frequency questionnaire.
