ABSTRACT
The calculation of measures of bivariate association among attribute values is often one of the –rst steps taken following the initial exploratory analysis of a data set. In this sense, measures of bivariate association form a sort of bridge between the exploratory and con-–rmatory stages of the analysis. When dealing with continuously valued data, two of the most commonly used measures of bivariate association are the Pearson product moment correlation coef–cient r and the Spearman rank correlation coef–cient rs. These statistics measure linear association between two quantities. The exploratory phase of the data analysis should, however, have established a general working hypothesis about the form of the relationships between variables, and the computation of the correlation coef–cient should be carried out if the results of this exploratory analysis indicate that a linear association is possible.
