One study of high research quality design reported that the Mifflin-St. Jeor equation accurately predicted RMR using actual body weight within +/- 10% of measured RMR in 70% of obese individuals. Of the remaining 30%, 9% were overestimations and 21% were underestimations. The individual error range was a maximum overestimate of 15% to a maximum underestimate of 20%.