Investigating the Significance of a Correlation Coefficient using Bootstrap Estimates

Resampling methods offers effective estimates of parameters and its asymptotic distribution. In this study, it is recommended to use the bootstrap method as an alternative to the classical and knife (one exclusion procedure) test statistics in evaluating the significance of the Pearson correlation coefficient by applying the bootstrap method to the simple linear regression model. This procedure provides an effective alternative to test the significance of the Pearson correlation coefficient. In the application, the model parameters, standard errors, Pearson coefficients of correlation, bias and % 95 confidence intervals belonging to bootstrap and jackknife methods in estimated with the help of a real data and the obtained results are interpreted. As a result, the test statistic obtained by the bootstrap method is proposed as an alternative to the classical and jackknife test statistics