Why self-reported food intake is systematically inaccurate.
What the evidence says
- People significantly underreport calorie intake when it is measured objectively — here over 14 days ”by indirect calorimetry and analysis of body composition” (corrected 2026-09-19 from ”objective measures like doubly labeled water”: the only text held for this study names calorimetry and body composition, not doubly labelled water)[1]
- Ten diet-resistant subjects (nine women and one man; the comparison group was 80) ”underreported their actual food intake by an average (+/- SD) of 47 +/- 16 percent” and ”overreported their physical activity by 51 +/- 75 percent” (corrected 2026-09-19: the 19% control figure is in no text this library holds for the study)[1]
- The underreporting is largely unconscious — subjects genuinely believe their reports[1]
- Mechanisms include selective memory (afternoon snacks were the single most underreported eating occasion), portion distortion (beverages underestimated by 30-46%, pasta 44%, pizza 32% — while high-energy-density snacks were overestimated), and forgetting incidental eating[2],[3]
- Portion estimation follows a “flat slope” pattern: large portions are underestimated while small portions are overestimated, with individual errors ranging from -88% to +900%[3]
- “Diet resistance” is typically explained by intake underestimation, not metabolic abnormality[1]