Objective: To compare energy expenditure ( EE) measured by doubly labeled water ( DLW) with other measures, both physical and based on subjective questionnaires. Design: A comparison of methods in a stratified sample of adult volunteers. Setting: The feeding behaviour suite ( FBS) at the Rowett Research Institute, Aberdeen. Subjects: A total of 59 subjects, stratified for age, sex and body mass index ( BMI). Interventions: EE was assessed by DLW ( validated using measurements of energy balance), heart rate monitor ( HRM), activity monitor ( Caltrac), 24- h physical activity diary ( PAD) and 7- day physical activity recall. Energy intake was assessed using covert ( investigator- weighed) food intake ( EI). Data were collected over a 12- day period of residence in the Rowett's FBS. Results: No methods correlated highly with physical activity assessed by DLW. Physical methods correlated more closely than did subjective recording. All methods ( except EI) significantly underestimated EE, estimated by DLW. There were no significant differences in association between methods and sex, age, BMI or fat- free mass. Conclusion: EE is difficult to measure precisely or accurately with current approaches but physical methods are slightly better than subjective accounts.