Traditionally data processing in applied biomechanics has relied on descriptive approaches. Although these are effective exploratory techniques, they may not provide an understanding of the interaction between the variables that describe the event across repeated sampling. Factor analysis is a statistical process that allows the researcher to extend prediction beyond a univariate model to a structural equation in which dependent variables are processed against latent factors. The purpose of this investigation was to examine the application of factor analysis in the reduction of input variables to minimize inter-subject variance in structural equation modelling in biomechanics and specifically competitive ski jumping. Applying a systematic method of data processing, variables that lacked robustness were omitted; while variables that maintained homogeneity of variance across the mid-flight phase of the jump were selected. Based on this analytical approach, the final model is less influenced by confounding from the implicit variance that arises when using sequences of random variables (heteroscedasticity) within a set of predictor variables. Therefore, the final model is expected to maintain the characteristics of homoscedasticity and minimize stochastic effects.
© Copyright 2013 International Journal of Sports Science and Coaching. Multi-Science Publishing. Alle Rechte vorbehalten.