The purpose of this study was to compare the normalization methods of ratio standards, allometry, and ANCOVA with knee extensor strength of older adults. The apparently healthy older volunteers were 71 men (mean +/- SD; age, 71 +/- 4 yr; body mass, 81 +/- 10 kg; height, 174 +/- 7 cm) and 77 women (71 +/- 4 yr, 65 +/- 8 kg, 160 +/- 5 cm. respectively). Strength was defined as peak torque (N.m-1) and measured with a Cybex II isokinetic dynamometer. Body composition was estimated with dual energy x-ray absorptiometry. With allometry, the body mass exponent (0.74) was not statistically different from theory (0.67). Body mass adjusted strengths were 34.7% (allometry), 32.0% (ANCOVA), and 29.4% (ratio standards) greater in older men than women. Allometry revealed that the bone-free lean tissue mass exponent was not different from ratio standard exponent of 1.0. After adjustment by bone-free lean tissue mass, strength in men remained 16.0% (allometry and ratio standards) higher than in women, but, strength differences between genders were eliminated with ANCOVA. The methods used to normalize strength yielded similar results with body mass but conflicting results with bone-free lean tissue mass.