Evaluation Of Vertical Jumping Performance Under The Perspective Of Principal Component Analysis

Introduction
Multiple regression analysis techniques have been used in the past [1] in an attempt to identify the most critical parameters determining vertical jumping (VJ) performance. The use of Principal Components Analysis (PCA), however, incorporates a large number of variables, highly correlated to each other, by a smaller number of computed factors [2]; thus, it has been shown that it could be a useful tool to evaluate vertical squat jumping [3]. The purpose of the present study was to explore whether PCA could be applied to examine a possible tendency of temporal or peak force dependency among groups of athletes executing different types of VJ.

Methods
11 male international-level jumpers (TF; 22.9±4.5yrs; 1.87±0.05m; 76.6±6.7kg), 10 male professional volleyball players (VB; 23.9±4.4yrs; 1.94±0.03m; 85.7±7.0kg) and 14 male professional soccer players (SO; 25.3±4.2yrs; 1.82±0.06m; 78.7±7.4kg) executed a) squat jumps with the use of arms (SQJA), b) countermovement jumps with (CMJF) and without the use of arms (CMJA) and c) drop jumps from 60cm without the use of arms (DJ60A) on an AMTI OR6-5-1 force plate (AMTI, Newton, MA). Data (sampling frequency: 500Hz) were stored in a 486 DX personal computer. A PCA with Varimax Rotation on the exerted dynamic parameters was conducted using the SPSS 10.0.1 software (SPSS Inc., Chicago, Il.). PCA scores were statistically analysed using one-way ANOVA with a Scheffe Post-Hoc test (p<.05) in order to reveal differences among groups.

Results
PCA results suggested the existence of two factors of principal components. The first rotated principal component was associated with the temporal characteristics of the VJ, which accounted for 56.0%, 60.3%, 57.2% and 55.3% of the variance of the SQJF, CMJA, CMJF and DJ60A force data, respectively. The second was associated with the peak force characteristics of the VJ, which accounted for 22.3%, 17.5%, 21.6% and 19.9% of the variance of the SQJF, CMJA, CMJF and DJ60A force data, respectively. Comparison of TF, VB and SO are presented in Table 1.

Conclusions
PCA scores derived from the participant subjects allowed a further insight to the relevance of parameters determining VJ performance, since different tendencies were observed. TF utilized a temporal dependency of jumping parameters, while SO seemed to rely on factors connected to force characteristics (Graph 1). This kind of examination not only could be a useful methodological approach for assessing VJ, but also for classifying athletes based on jumping characteristics. Finally, as shown by the present study, the PCA model used in the past [3] is also valid for countermovement jumps or jumps involving use of the arms.

References
[1] Aragon-Vargas l.F. & Gross M.M. (1997). J Appl. Biomech, 13, 45-65.
[2] Kleinbaum D.G. (1988). Applied regression analysis and other multivariable methods. Boston: Pws-Kent.
[3] Kollias I. et al. (2001). Res Q Exerc Sport, 72, 63-67.

(artigo completo com gráficos e tabelas no anexo)

Arquivo