Partial least squares-structural equation modeling (PLS-SEM) analysis of team success using R

Authors : Mehmet Türegün

Pages : 201-213

Doi:10.18826/useeabd.628653

View : 26 | Download : 11

Publication Date : 2019-12-15

Article Type : Research

Abstract :Aim:  A combination of football becoming highly commercialized, technological advances made, and increasing amounts of data becoming available has made it possible for researchers to conduct statistical analyses of the various aspects of the game with an ultimate focus on determining the key factors for team success. Methods:  This quasi-experimental study used an ex-post facto design to develop a model for team success. The sample consisted of 18 teams which played 306 matches in a 9-month long association football league format. A PLS-SEM path analysis was conducted using 11 latent variables.  Results:  Findings yielded a substantial overall model fit (GoF R2=0.811) for the measurement and structural models. The latent variables (LVs) of offence (β= 0.630, p< .001) and defence (β= 0.489, p<0.001) had statistically significant effects on the LV of success. The exogenous LVs offence and defence predicted 79.9% of the variability of the LV success and its manifest variables.  Conclusion:  The defensive ability of a team seemed just as important as the offensive ability for team success in football. This particular conclusion is well aligned with the outcome of various studies conducted by other researchers. For instance, Hughes & Churchill (2004) stated that in their study it appeared that defensive ability of teams to control the opposing team's movements had a significant effect on team success. 
Keywords : Team success, football, path analysis, modeling