Abstract
The purpose of this investigation was to compare reliability coefficients and standard errors of measurement (SEM) for absolute (peak torque in newtonmeters) and ratio (concentric/eccentric (C/E)) and knee flexion/extension (KF/KE) ratios—as percentages) data. Forty healthy volunteers (20 men and 20 women) completed knee extension and flexion tests on two occasions within a 2-week period. On each occasion, two concentric and two eccentric muscle actions, with a 5-second rest between muscle actions, were completed from 90° to 5° flexion at 60 deg/sec angular velocity. For each gender, absolute scores were characterized by higher interrepetition, interoccasion, and overall intraclass correlation coefficients and lower relative measurement error (SEM/SD) than were either C/E or KF/KE ratios. These results do not discount the utility of ratio data, but suggest that to maximize reliability, ratio data should be acquired using more repetitions and/or occasions than is the case when using absolute data.