Measuring Association for Interval-Level Data: Pearson’s Correlation Coefficient

Abstract

This chapter introduces the linear correlation coefficient, a widely used descriptive statistic that enables the researcher to describe the relationship between two interval-level measures. This situation is encountered often in criminal justice research. For example, researchers may want to establish whether number of prior arrests is related to age, education, or monthly income. similarly, it is common in criminal justice research to ask whether the severity of a sanction measured on an interval scale (e.g., number of years sentenced to imprisonment or amount of a fine) is related to such variables as the amount stolen in an offense or the number of prior arrests or convictions of a defendant. We also examine an alternative rank-order measure of association that may be used when the linear correlation coefficient will lead to misleading results.