Article Others

The relationship among skills/factors and their structure and wins/losses in games in basketball classes

This study aimed to extract the structure of plays/fouls of basketball games in university PE classes using MDS and to investigate the relationship between wins/losses and their structure using logistic regression analysis in the games. Targeted games were 20 games performed in university PE classes at F-university and the plays/factors observed were assessed by a 5-point scale. The items observed were 52 items chosen from “Passing,” “Dribbling,” “Shooting,” “Rebounding,” “Fast break,” “Cooperative plays,” “Violations,” and “Fouls.” After the principal component analysis was applied to the whole data, the items that did not show significant loadings were removed from the analysis, and a correlation matrix among only the items that had significant principle loadings was computed. Another principal component analysis was performed and the principal component scores for each skill/factor were computed. Nonmetric MDS was applied to the correlation matrix, a two-dimensional space was extracted, and the rotation of axes by Varimax criterion and clustering by hierarchical cluster analysis were applied to the two-dimensional space. Finally, logistic regression analysis to predict wins/losses was carried out using skills/factors including gender and evaluators to remove their influence as independent variables. As a result, the following results were obtained: 1) The two-dimensional space obtained was interpreted as “Individual skills and skills leading to fouls/violations” and “Individual – group skills.” Again, the four clusters obtained were also named “Ball-advancing individual skills,” “Under-basket individual skills,” “Cooperative skills,” and “Against-rule and unskilled plays.” 2) Although the likelihood ratio test of logistic regression analysis with all independent variables was not significant, the test of the model after the independent-variable selection was significant, in which “Shooting,” “Cooperative plays,” “Fouls,” and “Gender” were included. In the univariate analysis using correlation ratios, “Dribbling,” “Shooting,” and “Cooperative plays” were significant, but “Fouls” were not. It is thought that “Dribbling” lost its significant relationship with wins/losses when removing the influence of “Cooperative plays,” and insignificant “Fouls” in the univariate analysis showed an important impact on wins/losses because the fouls provided the change of offense and the risk of a losing score increased.

View more >>