The aim of the current study was to measure differences in reporting behaviour between a first occasion of 7 d food recording and a second occasion of 7 d food recording 12 weeks later, in a group of elderly men (n 17) and women (n 17). Half the group followed an exercise intervention. The mean age was 61 (SD 5) YEARS AND MEAN BMI WAS 26.2 (sd 3.8) kg/m2. Reported energy intake was compared with energy expenditure as calculated from measured BMR and physical activity assessed with a tri-axial accelerometer for movement registration. Total under-reporting was divided into undereating and under-recording. Undereating was calculated from the change in body mass over the recording week and the under-recording was measured using the water balance technique. In the first period, the total under-reporting was 21 % and increased to 27 % in the second period (P=0.03). In the first period there was no indication for subjects eating less during the recording week, however, in the second period subjects lost body mass during the food recording indicating undereating. The amount of under-recording was calculated at 21 % in the first period and 18 % in the second period of recording (P 0.28). During the second period subjects selectively under-reported their fat intake and over-reported their protein intake. In conclusion, repeated assessment of food intake caused a higher quantitative and a qualitative under-reporting of food intake. The effect of interventions (dietary or otherwise) on habitual food intake might be confounded by changes in food-reporting behaviour.