Objective: Large variations in maximal voluntary torque are reported in the literature during isometric plantarflexion contractions. We propose that these differences, which could reach 40 % across similar studies, could be explained by differences in the instructions provided, and notably by instructions as to favoring or not multi-joint contractions. Method: Sixteen participants were placed on an isokinetic ergometer in 3 different positions, supine, prone and seated, with the ankle in the neutral position, and instructed to create maximal force on the footplate by conforming to instructions that favored either isolated (ISOL) or multi-joint (ALL) isometric contractions. Torque, foot kinematics and the electromyographic activity of seven muscles of the lower limb have been recorded. Results: Joint torques were greater in ALL compared to ISOL (p < 0.05) with gains of 43.5 (25.4–170.6) %, 42.5 (1.4–194.6) % and 15.3 (9.3–71.9) % in the supine, prone and seated position, respectively [values are given as median (range)]. The results of this study suggested that forces created by muscles that do not span over the ankle joint significantly influenced the measured joint torque. Nevertheless, the observed gains in torque were associated with greater plantarflexor muscles activation, showing that the ISOL condition may have induced a form of inhibition of these muscles. Conclusions: The results of this study suggest that using isolated contractions, hence constrained testing protocols, cannot provide optimal conditions for MVC testing, notably for plantarflexor muscles, which seem to be extremely sensitive to such constrained conditions.