Computational protocol: Physique and Performance of Young Wheelchair Basketball Players in Relation with Classification

Similar protocols

Protocol publication

[…] Data were assessed for normality with the Shapiro–Wilk test and log-transformed where necessary. Descriptive statistics (mean and standard deviation) was computed for all variables using standard procedures. Means difference was assessed with two-sample t-test (two-tailed). To assess the relationships between demographic (age, WB experience, and assigned functional points), anthropometric and body composition, and performance (game-related statistics and field tests) variables, the Pearson’s product-moment correlation coefficient (r) and the Spearman’s rho (ρ) were used for continuous and categorical variables, respectively. The strength of the correlation coefficient was considered small (0.00–0.30), moderate (0.31–0.49), large (0.50–0.69), very large (0.70–0.89), and almost perfect for assessing relationships (0.90–1.00) as per Hopkins []. In order to minimize Type I error associated with multiplicity of correlations in the same dataset, the Benjamini and Hochberg False Discovery Rate procedure was used to get corrected P-value (Pc). For regression analysis, the participants were randomly assigned to two groups: a model development group (MD group, n = 35, 2/3 of sample) and a cross-validation group (CV group, n = 17, 1/3 of sample). The MD and CV groups were comprised of 13 and 6 Class A players, 10 and 6 Class B players, 6 and 2 Class C players, and 6 and 3 Class D players, respectively. In the MD group, separate stepwise multiple regression analyses (enter, F<0.05; remove, F>0.1) were run using demographic data, anthropometry and FM%, and field tests scores as independent variables to identify predictor(s) of individual of game-related statistics namely, TP, FT, and FG. Adjusted coefficients of determination (R2) and standard error of the estimate (SEE) were used to represent the goodness of the predictor model. Homoscedasticity of data was assessed by plotting the residuals of multiple regression analysis against the predicted values. The Durbin-Watson test, the variance inflation factor, the tolerance value, and the condition index were calculated to test collinearity. The developed regression models were then cross validated using the data from the CV group. Paired t-tests were performed to determine the difference between estimated game-related statistic and actual values. Reliability of data was assessed with the intraclass correlation coefficient (ICC; >0.75 = good reliability []), the Bland-Altman plot, and the standard error of measurement.One-Way ANOVA was used to evaluate differences in the demographic, anthropometric, body composition, and performance variables (game-related statistics and field tests) of the four functional ability Classes; in case of significance, post-hoc comparisons were carried out with Bonferroni’s correction for multiple comparisons. The Levene’s test was performed to validate the application of ANOVA. Cohen’s f (f) and Cohen’s f2 (f2) were used to calculate the effect size in the ANOVA and the regression analysis, respectively. 95% confidence intervals (CI) were calculated for Cohen’s f and Cohen’s f2 statistics. According to Cohen’s guidelines [], effect size values were interpreted as small (f = 0.1, f2 = 0.02), medium (f = 0.25, f2 = 0.15) and large (f = 0.4, f2 = 0.35) for ANOVA and for regression analysis effects, respectively. All analysis was performed with SPSS v. 16.0 (IBM Corp., Armonk, NewYork, USA). Post hoc statistical power of the sample was evaluated using G*Power Software 3.1 [] on the basis of the observed effect sizes. The alpha value was set at 0.05. […]

Pipeline specifications

Software tools SPSS, G*Power
Application Miscellaneous