ABSTRACT
Measurement of the association between variables is a fundamental component of statistical investigation. The correlation between two quantitative random variables quantifies association in a way that does not depend on units of measurement. Correlation and regression calculations are time-consuming for even moderate amounts of data but, luckily, the smallest microcomputer can handle the problem fairly straightforwardly. Although the instances involving at least one qualitative random variable are perfectly valid examples of association, the concept is at its most powerful when applied to two quantitative variables. Regression methods can also be applied to non-linear models when the background to the problem or inspection of the scattergram suggest a linearizing transformation. It is fundamental in scientific research to explore relationships of cause and effect and such investigations must first consider whether or not there exist associations among random variables.
