ABSTRACT
So as to set the stage for applying the commonly encountered Pearson's correlation coefficient, this chapter begins with an approach to assessing the validity of an assumption that a bivariate sample comes from a bivariate normal distribution from a graphical point of view. This is followed by explanations of the computational steps involved in using a sample to estimate Pearson's correlation coefficient, and perform the test of significance as well as testing against a non-zero hypothesized value. For bivariate samples not meeting the assumption of bivariate normality, Kendall's Tao and Spearman's Rho are discussed next, both from the exact as well as the normal approximation point of view. For non-numerical data (ordinal or nominal), the chi-square test of independence is illustrated, and then the chapter closes with a brief discussion and demonstrations of computing probability distributions for the distributions of Kendall's and Spearman's sample correlation coefficients.
