Structural Equation Modelling and Amos: Factor Analysis of Torque, Work and Power

The software AMOS, a second generation statistical method based on structural equation modelling (SEM), can address complex problems and interactions in sport and exercise science, and especially factor analytic research relevant to measurement theory in biomechanics, exercise physiology and motor skill learning. Sixty healthy and active participants (32 males; 28 females) were involved in the study located in an exercise laboratory. AMOS was applied as a method of confirmatory factor analysis (CFA) confirm a measurement model for torque-work and power based on isokinetic leg extension at isokinetic speeds of 60o s-1 (5 reps), 180os-1 (15 reps) and 300os-1 (15 reps) using leg extension/flexion with CYBEX technology and measures of power using vertical jump concepts of height, flight time, contact time and 10m sprint acceleration. The hypothesis was that many dependent variables are only measuring a small set of underlying factors or latent constructs. The results indicated that the two factor model was partially supported using SEM goodness of fit indices, however more substantively by arguments based on construct, convergent and discriminant validity. The discussion promotes CFA-SEM as methods to understand more completely sports specific fitness tests based on multiple measurements and measurement parsimony.