Under-reporting of food and drink intake is common and represents a major problem for dietary studies. People may:

• Fail to report items they have actually consumed
• Report what they perceive to be a "more healthy" intake than normally consumed, in an attempt to avoid criticism
• Fail to follow their normal diet while being studied

Under-reporting is more common in those who are overweight or trying to control their weight than in normal weight individuals, but does occur in normal weight individuals. It is more common in women than in men. It occurs in children as well as adults. The prevalence of under-reporting is increasing and has been reported to occur in as many as 40% of the population.

Under-reporting introduces significant bias into studies. Under-reporters can be identified either by:

• The doubly-labelled water method to measure energy expenditure (used to validate energy intake of individuals in energy balance).
• Comparison of energy intake with calculated Basal Metabolic Rate (BMR)

If under-reporters are identified and then excluded from statistical analyses, this also introduces bias as the remaining subjects may no longer be representative of the population from which they were selected. It has been suggested that bias resulting from under-reporting can be reduced by adjusting for energy intake.

The ideal would be to reduce the occurence of under-reporting among those being surveyed. Studies have shown that this is possible (see Fehily, 2007).

Reference:
Fehily A. Dietary assessment. In: Manual of Dietetic Practice, Fourth Edition. (Eds B Thomas and J Bishop). Oxford: Blackwell Scientific Publications, 2007, p.32-40.