The purpose of this study was to compare the ability of various body-composition assessment techniques to detect changes in soft tissue in older, weight-stable women (50-70 y of age) completing a 1-y randomized, controlled trial of progressive resistance training. The intervention group (n = 20) performed high-intensity strength-training 2 d/wk with five different exercises; the control group (n = 19) was untreated. Hydrostatic weighing, 24-h urinary creatinine, computed tomography of thigh sections, total body potassium, and tritium dilution techniques were used to measure increases in total fat-free mass (FFM) and the muscle and water components of FFM. A decrease in fat mass (by hydrostatic weighing) was seen in the strength-trained women compared with the control subjects (P - 0.01-0.0001). Anthropometry, bioelectric impedance, dual-energy X-ray absorptiometry, and total body nitrogen and carbon did not measure any significant change in soft tissue. The choice of a body-composition technique is important when designing a study expected to affect soft tissue, because not all techniques available are precise enough to detect small changes.