ABSTRACT
Correlation measures the strength of the linear association between two continuous variables. Both Pearson and Spearman rank (nonparametric counterpart) correlation coefficients are introduced, as well as how to use scatter plots to get a sense of the linear association between two variables. The scatter plots help readers to understand how the calculated correlation coefficients look when graphed. The chapter provides users with a recap on creating scatter plots using a statistical software program (SAS and Stata). Correlation is also discussed in a hypothesis testing framework and is related to the basic problem of statistics, using the sample correlation to make inferences about the correlation in the population from which the sample was obtained. Specifically, is the population correlation significantly different from zero? Conveyed differently, is there a linear association between the two variables? Although formulas are provided for the calculation of correlation coefficients and conducting hypothesis tests, users are encouraged to use statistical software; datasets (in SAS and Stata formats) for practice problems are provided on the book website.
