The correlation coefficient is greatly influenced by outliers, therefore, a method called Spearman’s rho ranks the values to remove this effect. It is also useful for ordinal data where the observations have no actual numerical value, but they can be ranked from least to greatest. The observations for each variable are ranked from least to greatest independently of one another, but the pairs remain together. For example, one observation may have the lowest x value and the tenth lowest y value. In this case, the ordered pair is (1,10). These pairs are plotted on a scatter plot, and the correlation coefficient, Spearman’s rank correlation is calculated by
Now the methods for hypothesis testing of the coefficient described for Pearson coefficients can be used.