Conventional Ratio Scaling Overcorrects for Body Mass When Evaluating Isokinetic Knee Strength in Elite Female Soccer Players

Isokinetic knee strength is commonly normalized to body mass, although this assumes that torque increases in direct proportion to mass. This study tested whether conventional ratio scaling removed body-mass dependence from concentric knee-extensor and knee-flexor peak torque in elite female soccer players. Sixty-seven preseason assessments from 46 players were analyzed. Bilateral mean peak torque was calculated at 60°·s−1 and 300°·s−1, and log-linear models estimated the exponent b in torque = a·massb using player-clustered standard errors. Player-mean sensitivity analyses and leave-one-player-out internal cross-validation were also performed. Body mass was positively associated with absolute torque but negatively associated with torque divided by body mass. The estimated exponents were 0.562 and 0.551 for quadriceps and hamstring torque at 60°·s−1 and 0.349 and 0.374 at 300°·s−1. All confidence intervals excluded 1.00. Player-mean estimates were similar, and cross-validated residuals showed little remaining linear association with body mass within this cohort. Knee torque increased with body mass, but less than proportionally. Direct division by body mass therefore overcorrected for size and tended to favor lighter players. Absolute torque should continue to be reported, while the sample-specific exponents require external validation before broader application.