Because of the need for efficient, consistent strength measurements, the test-retest reliability of concentric, isometric, and eccentric strength; concentric work; and concentric power was determined in older women without a familiarization session. The reliability of measures derived from a single peak score were compared with those derived from an averaged score. On 2 occasions 25 older women with a mean age of 72 +/- 6 years performed 3 submaximal knee extensions and 5 maximal contractions on an isokinetic dynamometer at 90 degrees/s (CON), 0 degrees/s, and -90 degrees/s on both lower limbs. Statistical analyses for peak and averaged values (best 3 contractions of 5) exhibited good relative reliability (ICCs > .88), except for CON power. Typical error as a coefficient of variation and ratio limits of agreement for peak and averaged score values were larger than desired, with CON power scores demonstrating unacceptable error ranges. Although relative reliability of this 1-session assessment protocol was acceptable, further research is needed to determine whether additional practice trials could enhance absolute reliability.