Abstract
Fisher's transformation to determine confidence intervals for Pearson's product-moment correlation coefficient. In this study we test the accuracy of Fisher's Z transformation for estimating confidence intervals for Pearson's product-moment correlation coefficient with several correlation values and sample sizes and under several conditions of normality and non normality of residuals. In conclusion, under normality of residuals and given a sufficiently large number of subjects, the determination of confidence intervals by means of such transformation is adequate.