This study assessed a simple factorial method to predict energy expenditure (EE) in confined and free-living subjects. Thirteen subjects (seven male, 6 female) were studied on four occasions each. Measurements included BMR, 24-h EE by continuous respirometry, 4-d records of intake and activities, body weight, and urine collections. Agreement between measured and predicted 24-h EE was within +/- 2% for group results and +/- 10% for most individuals. Individual agreement was improved to within +/- 5% by using measured rather than predicted BMR. Habitual energy intake was 7.3% greater (p less than 0.05) than predicted EE in males and 5.1% less (NS) in females. Validatory data suggested overrecording of intake in males and underrecording in females. Large individual differences between habitual intake and predicted EE were observed but were significantly (p less than 0.01) correlated with both weight change and protein balance. Our simple factorial method provided acceptable estimates of EE in the subjects studied.