The New England Journal of medicine did a study with "doubly labeled water" which made it possible to determine _exactly_ how many calories a person ate. In a

The New England Journal of medicine did a study with "doubly labeled water" which made it possible to determine _exactly_ how many calories a person ate. In a study of people who claimed to be "diet resistant" (some of whom were taking medication based on what they reported to their doctors) it had been determined that each and every one of them were underreporting their calorie intake and overreporting their exercise to the doctors. Here's the study (PDF) http://www.nejm.org/doi/pdf/10.1056/NEJM199212313272701 http://www.nejm.org/doi/pdf/10.1056/NEJM199212313272701 It's been duplicated again and again. The last paragraph says: "In conclusion, all the obese subjects we studied who had a history of self-reported diet resistance had appropriate energy expenditure, but they misreported their actual food intake and physical activity." It's sounds shocking to me, but people would rather lie to their doctors, and get medication they don't need, than count calories accurately.