This study aimed to examine the validity of mechanical variables obtained by a linear velocity transducer from the unconstrained and constrained squat jump (SJ). Twenty-three men were tested on the unconstrained SJ and the SJ constrained by a Smith machine. Maximum values of force, velocity, and power were simultaneously recorded both by a linear velocity transducer attached to a bar of mass of 17, 30, 45, 60, and 75 kg and by a force plate. Linear velocity transducer generally overestimated the outcomes measured as compared to the force plate, particularly in unconstrained SJ. Bland-Altman plots revealed that heteroscedasticity of errors was mainly observed for velocity variables (r 2 = .26-.58) where the differences were negatively associated with the load magnitude. However, exceptionally high correlations were observed between the same outcomes recorded with the 2 methods in both unconstrained (median r = .89 [.71-.95]) and constrained SJ (r = .90 [.65-.95]). Although the systematic and proportional bias needs to be acknowledged, the high correlations between the variables obtained by 2 methods suggest that the linear velocity transducer could provide valid values of the force, velocity, and power outputs from both unconstrained and constrained SJ.
Keywords: Smith machine; force; free-weight; power; velocity.