Many nutrition studies rely on self-reported dietary intakes. Surveys ask participants to record everything they eat for a specific period, recall what they ate recently, or indicate what foods (in what amounts) they typically eat. The information these surveys provide is used to determine what dietary components or habits are associated with particular health measures or outcomes. If the self-reported diet information is inaccurate, it will impact study results.
The American Journal of Clinical Nutrition recently published a study that sought to estimate the prevalence of under- and overreporting on food intake surveys. The study compared answers to three common self-reporting methods (24-hour recall, four-day food records, and food-frequency questionnaires) to laboratory measurements of actual levels of nutrients and calorie intake. The study included 530 men and 545 women aged 50 to 74 years. The authors concluded that all three self-reporting methods are useful, although food-frequency questionnaires resulted in the largest variances from lab results. Energy intake (calories) was consistently underreported (by 15 to 34 percent) in all methods, especially when using food-frequency questionnaires and among obese individuals. Reported intake of other nutrients was also more likely to be underreported than overreported.
There is no perfect way for researchers to determine what people eat over time. Consumers should be wary of media reports touting the results of one particular study (particularly if that study didn’t involve large numbers of people). The answers to nutrition questions lie in looking at a large body of research over time.